{ "text": " Hello! A new model has sparked widespread discussion over the last couple of days, named JF. A former OpenAI researcher spent two years developing it after leaving, creating a new model trained in a way that differs from the past. I think its features are very distinct and powerful. For specific tasks, it can run at a fraction of the cost and hundreds of times the speed of traditional models, with very low hallucination rates, generating the answers you need, which is highly significant.它是重要的,因為它選了完全不同的一路由我們現代的AI,它是在原來的N-Token Prediction。當大家是比拼了強硬的解決,硬的解決,和能力的解決方法,它其實是很小的地步。我們不需要等待太多的信息,或多多的解決,我們只需要一個比較的測試,或是個比例的不同的結果。我给你一个例子例如,在一个公司的行业,如果我们使用了AI的相机,我们必须要进入到的模特,等待在模特上的模特,这些时间需要几秒,对吗?这一终的效果会给你,你会在哪个部份接触,谁要跟他在谈论,或者是指出的优质是低的,需要更加注意力。它给你一个样的效果。在实际上,用这个新模特,它可以非常快速,这几个百分钟的几个百分钟的几个百分钟这几个百分钟的几个百分钟这几个百分钟的几个百分钟the billing department,estimatesthe customer churn rate,and assessesthe likelihoodofneedinghuman intervention.Once theseassessmentsare made,I canassignthe tasktoaspecificdepartmentorouteittoahumanagent,allowingme tomakethecorrespondingadjustmentsanddecisions.So itsefficiency andcosteffectivenessareincrediblyimpressive.Itisjustextremely,extremelyfast.Even initslivedemonstrationofstructuredoutput,youcanclearlyseethatitgeneratesresults但是传统的 models也会一直在用评论词为文字的评论所以这两个技术的 approaches其实完全不同当 the founder released this video他说了一篇,他说了一篇,我们的认识,其实是一篇的答案合适的答案不太不太确定,我们喜欢他们,对吗?例如,他们可能会更多 empathetic或是更多理解我们的但是,他們的問題是他們的解答問題人們不可能會不喜歡,甚至是有個方法的方法。這帶來更加盡量的信息和 consistency。但是,他們的解答案是叫做RLCD,名為 Reinforcement Learning from Calibrated Decisions。它指的是,在 Reinforcement Learning,我實在是在調整的解答案是的解答案是一個解答案。解答案是每次它提供答案,它提供一個 probability。A confidence scoreFor instance, with this billing inquirythe probability that I should route it to the billing department is about 90%But if the customer also mentioned other issuesI believe there's a certain probability it should be prioritized for other departmentsSo I might assign a 10% or 20% probability to those as wellAt this point, you can set a thresholdFor example, if the churn risk is very highI can have a human intervene directly如果它是低,我可以让它 AI model handle the response.所以,它帮助我们去解决哪个性能量的权量使用来解决于不同的情况下。所以,我相信这个模型会将将其进行我们的生活生活和很多不同的专业。以前我们曾经用了一枪创造的创造,或者我们不需要太多的时间。但这些随意的随意的选择是我们的选择。但是在这个新的模式中,我们的选择和认证能更加更加更加适合。如果我们说我们的选择这个进入我们的页务,我们肯定能够用这个模式,因为现在,我们的页务员工程,他们其实会有很多认证和评价的评价。例如,如果我们的收入的收入我们的预期,我们还是需要这些认证,这会常常需要一个人的路程式的方法。但是如果我们有一个模式的模式,我们可以定义它的质量从开始即使是80%,60%,40%的距离,我们只需要在一个很低的40%的距离。如果它基本上是在结构的质量,我们不需要保持人的质量,我们不需要保持人的质量。所以,在这个将来的更多的模式,它可以更加大幅并且在这些模式上这个模式其实只发了一个新版本没有提供的技术和计程度我觉得这个方法非常值得考虑它使用一个系统一式的方法是快思考 Daniel Kahneman的作词说一大部分人的认识不在实际上的认识我们立刻的认识我们立刻的认识例如,选择哪个衣服很自然的衣服我只穿着衣服and when I'm going to work I might wear a different pair.So this judgment is actually very intuitive.It's the same for AI models.You don't actually need to generate or think through every single judgmentor use chain of thought reasoning.I can use a smaller model or a model designed for fast thinkingwith parallel computingto accomplish our tasks and goals very efficiently.So I think it might really make its way into our society soonbecause it's cheap, right?而且是最高的,因為它是很高的,在0.042萬元的價值,比 GPT-4的價值,可能是10萬元。價值是這麼高,他們現在不在賣現在的價值。所以它可以使用在一大多數的評測 scenarios。我們看到每個小型的手法可能有一個決定的模式背後。以前,一名 classifier would have required someone to spend ages training it, right?我們以前要有的要求,但現在是沒有必要的現在,這 model可以幫我們做了一切的評判,決定定 specific outcomes based on the context這代表,這代表,已經融化了更多的技術和考驗我們會看到這個功能伸展開更多的使用是的,我认为这个是非常有趣的新方式,一遍完全不同的方法使用力我們之前說明,未來可能會有小型的模式我們只需要有小型的模式提供的模式而是用販的模式每個都市場的模式我們最近的書我們最近的書也有同一個同意成為我們目標的最適合的方法是完全不平的所以我們想分享這些 JB的模式They've opened their waitlist, so if you'd like to try them out, feel free to join.Have a great weekend. Bye!", "segments": [ { "id": 0, "seek": 0, "start": 0.0, "end": 5.6000000000000005, "text": " Hello! A new model has sparked widespread discussion over the last couple of days, named JF.", "tokens": [ 50365, 2425, 0, 316, 777, 2316, 575, 39653, 22679, 5017, 670, 264, 1036, 1916, 295, 1708, 11, 4926, 40951, 13, 50645 ], "temperature": 0.0, "avg_logprob": -0.12579556046245255, "compression_ratio": 1.6237623762376239, "no_speech_prob": 5.254626733730561e-10 }, { "id": 1, "seek": 0, "start": 5.6000000000000005, "end": 14.24, "text": " A former OpenAI researcher spent two years developing it after leaving, creating a new model trained in a way that differs from the past.", "tokens": [ 50645, 316, 5819, 7238, 48698, 21751, 4418, 732, 924, 6416, 309, 934, 5012, 11, 4084, 257, 777, 2316, 8895, 294, 257, 636, 300, 37761, 490, 264, 1791, 13, 51077 ], "temperature": 0.0, "avg_logprob": -0.12579556046245255, "compression_ratio": 1.6237623762376239, "no_speech_prob": 5.254626733730561e-10 }, { "id": 2, "seek": 0, "start": 14.24, "end": 23.6, "text": " I think its features are very distinct and powerful. For specific tasks, it can run at a fraction of the cost and hundreds of times the speed of traditional models,", "tokens": [ 51077, 286, 519, 1080, 4122, 366, 588, 10644, 293, 4005, 13, 1171, 2685, 9608, 11, 309, 393, 1190, 412, 257, 14135, 295, 264, 2063, 293, 6779, 295, 1413, 264, 3073, 295, 5164, 5245, 11, 51545 ], "temperature": 0.0, "avg_logprob": -0.12579556046245255, "compression_ratio": 1.6237623762376239, "no_speech_prob": 5.254626733730561e-10 }, { "id": 3, "seek": 0, "start": 23.6, "end": 29.44, "text": " with very low hallucination rates, generating the answers you need, which is highly significant.", "tokens": [ 51545, 365, 588, 2295, 35212, 2486, 6846, 11, 17746, 264, 6338, 291, 643, 11, 597, 307, 5405, 4776, 13, 51837 ], "temperature": 0.0, "avg_logprob": -0.12579556046245255, "compression_ratio": 1.6237623762376239, "no_speech_prob": 5.254626733730561e-10 }, { "id": 4, "seek": 2944, "start": 29.44, "end": 34.32, "text": "它是重要的,因為它選了完全不同的一路由我們現代的AI,", "tokens": [ 50365, 11284, 1541, 24928, 1546, 11, 11471, 11284, 21625, 2289, 37100, 47123, 1546, 2257, 24658, 23786, 5884, 9581, 19105, 1546, 48698, 11, 50609 ], "temperature": 0.0, "avg_logprob": -1.2908005714416504, "compression_ratio": 1.2627737226277371, "no_speech_prob": 1.0279687257669368e-12 }, { "id": 5, "seek": 2944, "start": 34.32, "end": 36.32, "text": "它是在原來的N-Token Prediction。", "tokens": [ 50609, 11284, 1541, 3581, 19683, 3763, 1546, 45, 12, 51, 8406, 32969, 4105, 1543, 50709 ], "temperature": 0.0, "avg_logprob": -1.2908005714416504, "compression_ratio": 1.2627737226277371, "no_speech_prob": 1.0279687257669368e-12 }, { "id": 6, "seek": 2944, "start": 36.32, "end": 39.84, "text": "當大家是比拼了強硬的解決,", "tokens": [ 50709, 13118, 6868, 1541, 11706, 6852, 120, 2289, 22752, 38114, 105, 1546, 17278, 33540, 11, 50885 ], "temperature": 0.0, "avg_logprob": -1.2908005714416504, "compression_ratio": 1.2627737226277371, "no_speech_prob": 1.0279687257669368e-12 }, { "id": 7, "seek": 2944, "start": 39.84, "end": 44.480000000000004, "text": "硬的解決,和能力的解決方法,", "tokens": [ 50885, 38114, 105, 1546, 17278, 33540, 11, 12565, 8225, 13486, 1546, 17278, 33540, 9249, 11148, 11, 51117 ], "temperature": 0.0, "avg_logprob": -1.2908005714416504, "compression_ratio": 1.2627737226277371, "no_speech_prob": 1.0279687257669368e-12 }, { "id": 8, "seek": 2944, "start": 44.480000000000004, "end": 47.36, "text": "它其實是很小的地步。", "tokens": [ 51117, 11284, 14139, 1541, 4563, 7322, 1546, 10928, 31429, 1543, 51261 ], "temperature": 0.0, "avg_logprob": -1.2908005714416504, "compression_ratio": 1.2627737226277371, "no_speech_prob": 1.0279687257669368e-12 }, { "id": 9, "seek": 2944, "start": 47.36, "end": 49.6, "text": "我們不需要等待太多的信息,", "tokens": [ 51261, 5884, 1960, 35748, 10187, 18390, 9455, 6392, 1546, 17665, 26460, 11, 51373 ], "temperature": 0.0, "avg_logprob": -1.2908005714416504, "compression_ratio": 1.2627737226277371, "no_speech_prob": 1.0279687257669368e-12 }, { "id": 10, "seek": 2944, "start": 49.6, "end": 54.96, "text": "或多多的解決,我們只需要一個比較的測試,", "tokens": [ 51373, 19780, 6392, 6392, 1546, 17278, 33540, 11, 5884, 14003, 35748, 8990, 25174, 1546, 9592, 105, 22099, 11, 51641 ], "temperature": 0.0, "avg_logprob": -1.2908005714416504, "compression_ratio": 1.2627737226277371, "no_speech_prob": 1.0279687257669368e-12 }, { "id": 11, "seek": 2944, "start": 54.96, "end": 57.040000000000006, "text": "或是個比例的不同的結果。", "tokens": [ 51641, 19780, 1541, 3338, 11706, 17797, 1546, 47123, 1546, 35181, 1543, 51745 ], "temperature": 0.0, "avg_logprob": -1.2908005714416504, "compression_ratio": 1.2627737226277371, "no_speech_prob": 1.0279687257669368e-12 }, { "id": 12, "seek": 5704, "start": 57.04, "end": 59.6, "text": "我给你一个例子", "tokens": [ 50365, 1654, 23197, 2166, 20182, 17797, 7626, 50493 ], "temperature": 0.0, "avg_logprob": -1.208091239252034, "compression_ratio": 1.3513513513513513, "no_speech_prob": 8.533253964548848e-12 }, { "id": 13, "seek": 5704, "start": 59.6, "end": 63.18, "text": "例如,在一个公司的行业,", "tokens": [ 50493, 17797, 8238, 11, 3581, 20182, 13545, 32981, 1546, 8082, 940, 248, 11, 50672 ], "temperature": 0.0, "avg_logprob": -1.208091239252034, "compression_ratio": 1.3513513513513513, "no_speech_prob": 8.533253964548848e-12 }, { "id": 14, "seek": 5704, "start": 63.18, "end": 65.48, "text": "如果我们使用了AI的相机,", "tokens": [ 50672, 13119, 15003, 22982, 9254, 2289, 48698, 1546, 15106, 37960, 11, 50787 ], "temperature": 0.0, "avg_logprob": -1.208091239252034, "compression_ratio": 1.3513513513513513, "no_speech_prob": 8.533253964548848e-12 }, { "id": 15, "seek": 5704, "start": 65.48, "end": 67.14, "text": "我们必须要进入到的模特,", "tokens": [ 50787, 15003, 28531, 10178, 119, 4275, 36700, 14028, 4511, 1546, 41908, 17682, 11, 50870 ], "temperature": 0.0, "avg_logprob": -1.208091239252034, "compression_ratio": 1.3513513513513513, "no_speech_prob": 8.533253964548848e-12 }, { "id": 16, "seek": 5704, "start": 67.14, "end": 68.88, "text": "等待在模特上的模特,", "tokens": [ 50870, 10187, 18390, 3581, 41908, 17682, 5708, 1546, 41908, 17682, 11, 50957 ], "temperature": 0.0, "avg_logprob": -1.208091239252034, "compression_ratio": 1.3513513513513513, "no_speech_prob": 8.533253964548848e-12 }, { "id": 17, "seek": 5704, "start": 68.88, "end": 70.98, "text": "这些时间需要几秒,对吗?", "tokens": [ 50957, 5562, 13824, 44848, 35748, 6336, 254, 20956, 11, 8713, 14769, 30, 51062 ], "temperature": 0.0, "avg_logprob": -1.208091239252034, "compression_ratio": 1.3513513513513513, "no_speech_prob": 8.533253964548848e-12 }, { "id": 18, "seek": 5704, "start": 70.98, "end": 72.1, "text": "这一终的效果会给你,", "tokens": [ 51062, 5562, 2257, 10115, 230, 1546, 43076, 9319, 12949, 23197, 2166, 11, 51118 ], "temperature": 0.0, "avg_logprob": -1.208091239252034, "compression_ratio": 1.3513513513513513, "no_speech_prob": 8.533253964548848e-12 }, { "id": 19, "seek": 5704, "start": 72.1, "end": 74.1, "text": "你会在哪个部份接触,", "tokens": [ 51118, 2166, 12949, 3581, 17028, 7549, 13470, 36266, 14468, 6758, 99, 11, 51218 ], "temperature": 0.0, "avg_logprob": -1.208091239252034, "compression_ratio": 1.3513513513513513, "no_speech_prob": 8.533253964548848e-12 }, { "id": 20, "seek": 5704, "start": 74.1, "end": 75.2, "text": "谁要跟他在谈论,", "tokens": [ 51218, 33556, 4275, 9678, 5000, 3581, 8897, 230, 7422, 118, 11, 51273 ], "temperature": 0.0, "avg_logprob": -1.208091239252034, "compression_ratio": 1.3513513513513513, "no_speech_prob": 8.533253964548848e-12 }, { "id": 21, "seek": 5704, "start": 75.2, "end": 78.1, "text": "或者是指出的优质是低的,", "tokens": [ 51273, 31148, 1541, 25922, 7781, 1546, 7384, 246, 18464, 101, 1541, 41377, 1546, 11, 51418 ], "temperature": 0.0, "avg_logprob": -1.208091239252034, "compression_ratio": 1.3513513513513513, "no_speech_prob": 8.533253964548848e-12 }, { "id": 22, "seek": 5704, "start": 78.1, "end": 79.38, "text": "需要更加注意力。", "tokens": [ 51418, 35748, 19002, 9990, 40794, 13486, 1543, 51482 ], "temperature": 0.0, "avg_logprob": -1.208091239252034, "compression_ratio": 1.3513513513513513, "no_speech_prob": 8.533253964548848e-12 }, { "id": 23, "seek": 5704, "start": 79.38, "end": 80.88, "text": "它给你一个样的效果。", "tokens": [ 51482, 11284, 23197, 2166, 20182, 14496, 1546, 43076, 9319, 1543, 51557 ], "temperature": 0.0, "avg_logprob": -1.208091239252034, "compression_ratio": 1.3513513513513513, "no_speech_prob": 8.533253964548848e-12 }, { "id": 24, "seek": 5704, "start": 80.88, "end": 83.1, "text": "在实际上,用这个新模特,", "tokens": [ 51557, 3581, 24726, 8842, 227, 5708, 11, 9254, 15368, 12560, 41908, 17682, 11, 51668 ], "temperature": 0.0, "avg_logprob": -1.208091239252034, "compression_ratio": 1.3513513513513513, "no_speech_prob": 8.533253964548848e-12 }, { "id": 25, "seek": 5704, "start": 83.1, "end": 84.9, "text": "它可以非常快速,", "tokens": [ 51668, 11284, 6723, 14392, 10251, 31217, 11, 51758 ], "temperature": 0.0, "avg_logprob": -1.208091239252034, "compression_ratio": 1.3513513513513513, "no_speech_prob": 8.533253964548848e-12 }, { "id": 26, "seek": 8490, "start": 84.9, "end": 88.9, "text": "这几个百分钟的几个百分钟的几个百分钟", "tokens": [ 50365, 5562, 6336, 254, 7549, 31906, 6627, 50064, 1546, 6336, 254, 7549, 31906, 6627, 50064, 1546, 6336, 254, 7549, 31906, 6627, 50064, 50565 ], "temperature": 0.0, "avg_logprob": -0.2638493516710069, "compression_ratio": 8.017241379310345, "no_speech_prob": 9.544575416478573e-13 }, { "id": 27, "seek": 8490, "start": 88.9, "end": 90.9, "text": "这几个百分钟的几个百分钟", "tokens": [ 50565, 5562, 6336, 254, 7549, 31906, 6627, 50064, 1546, 6336, 254, 7549, 31906, 6627, 50064, 50665 ], "temperature": 0.0, "avg_logprob": -0.2638493516710069, "compression_ratio": 8.017241379310345, "no_speech_prob": 9.544575416478573e-13 }, { "id": 28, "seek": 8490, "start": 90.9, "end": 92.9, "text": "这几个百分钟的几个百分钟", "tokens": [ 50665, 5562, 6336, 254, 7549, 31906, 6627, 50064, 1546, 6336, 254, 7549, 31906, 6627, 50064, 50765 ], "temperature": 0.0, "avg_logprob": -0.2638493516710069, "compression_ratio": 8.017241379310345, "no_speech_prob": 9.544575416478573e-13 }, { "id": 29, "seek": 9290, "start": 92.9, "end": 98.18, "text": "the billing department,estimatesthe customer churn rate,and assessesthe likelihoodofneeding", "tokens": [ 50365, 3322, 35618, 5882, 11, 377, 332, 1024, 3322, 5474, 417, 925, 3314, 11, 474, 5877, 279, 3322, 22119, 2670, 716, 9794, 50629 ], "temperature": 0.0, "avg_logprob": -0.3331916168445849, "compression_ratio": 1.6029411764705883, "no_speech_prob": 1.2893640148614605e-12 }, { "id": 30, "seek": 9290, "start": 98.18, "end": 101.58000000000001, "text": "human intervention.", "tokens": [ 50629, 18796, 13176, 13, 50799 ], "temperature": 0.0, "avg_logprob": -0.3331916168445849, "compression_ratio": 1.6029411764705883, "no_speech_prob": 1.2893640148614605e-12 }, { "id": 31, "seek": 9290, "start": 101.58000000000001, "end": 107.9, "text": "Once theseassessmentsare made,I canassignthe tasktoaspecificdepartmentorouteittoahumanagent,", "tokens": [ 50799, 47062, 613, 640, 442, 1117, 543, 1027, 11, 40, 393, 640, 788, 3322, 5633, 1353, 64, 29258, 19929, 3033, 284, 14040, 270, 1353, 64, 18796, 559, 317, 11, 51115 ], "temperature": 0.0, "avg_logprob": -0.3331916168445849, "compression_ratio": 1.6029411764705883, "no_speech_prob": 1.2893640148614605e-12 }, { "id": 32, "seek": 9290, "start": 107.9, "end": 114.52000000000001, "text": "allowingme tomakethecorrespondingadjustmentsanddecisions.So itsefficiency andcosteffectivenessareincrediblyimpressive.It", "tokens": [ 51115, 336, 9637, 1398, 281, 38705, 3322, 19558, 6663, 278, 345, 3424, 1117, 474, 42821, 4252, 13, 6455, 1080, 68, 49086, 293, 27718, 46168, 8477, 543, 4647, 986, 3545, 8814, 22733, 13, 3522, 51446 ], "temperature": 0.0, "avg_logprob": -0.3331916168445849, "compression_ratio": 1.6029411764705883, "no_speech_prob": 1.2893640148614605e-12 }, { "id": 33, "seek": 9290, "start": 114.52000000000001, "end": 122.74000000000001, "text": "isjustextremely,extremelyfast.Even initslivedemonstrationofstructuredoutput,youcanclearlyseethatitgeneratesresults", "tokens": [ 51446, 271, 3424, 3828, 2579, 736, 11, 3828, 2579, 736, 7011, 13, 25931, 294, 1208, 45273, 10730, 4068, 2405, 2670, 372, 46847, 346, 2582, 11, 5616, 7035, 43679, 356, 17109, 6780, 270, 21848, 1024, 495, 33361, 51857 ], "temperature": 0.0, "avg_logprob": -0.3331916168445849, "compression_ratio": 1.6029411764705883, "no_speech_prob": 1.2893640148614605e-12 }, { "id": 34, "seek": 12274, "start": 122.74, "end": 125.08, "text": "但是传统的 models", "tokens": [ 50365, 11189, 7384, 254, 10115, 253, 1546, 5245, 50482 ], "temperature": 0.0, "avg_logprob": -1.2080620430611275, "compression_ratio": 1.2697841726618706, "no_speech_prob": 7.549700881820387e-13 }, { "id": 35, "seek": 12274, "start": 125.08, "end": 126.32, "text": "也会一直在用评论", "tokens": [ 50482, 6404, 12949, 34448, 3581, 9254, 5233, 226, 7422, 118, 50544 ], "temperature": 0.0, "avg_logprob": -1.2080620430611275, "compression_ratio": 1.2697841726618706, "no_speech_prob": 7.549700881820387e-13 }, { "id": 36, "seek": 12274, "start": 126.32, "end": 128.18, "text": "词为文字的评论", "tokens": [ 50544, 5233, 235, 13992, 17174, 22381, 1546, 5233, 226, 7422, 118, 50637 ], "temperature": 0.0, "avg_logprob": -1.2080620430611275, "compression_ratio": 1.2697841726618706, "no_speech_prob": 7.549700881820387e-13 }, { "id": 37, "seek": 12274, "start": 128.18, "end": 130.34, "text": "所以这两个技术的 approaches", "tokens": [ 50637, 7239, 5562, 36257, 7549, 32502, 1474, 107, 1546, 11587, 50745 ], "temperature": 0.0, "avg_logprob": -1.2080620430611275, "compression_ratio": 1.2697841726618706, "no_speech_prob": 7.549700881820387e-13 }, { "id": 38, "seek": 12274, "start": 130.34, "end": 132.22, "text": "其实完全不同", "tokens": [ 50745, 41646, 37100, 47123, 50839 ], "temperature": 0.0, "avg_logprob": -1.2080620430611275, "compression_ratio": 1.2697841726618706, "no_speech_prob": 7.549700881820387e-13 }, { "id": 39, "seek": 12274, "start": 132.22, "end": 133.96, "text": "当 the founder released this video", "tokens": [ 50839, 16233, 264, 14917, 4736, 341, 960, 50926 ], "temperature": 0.0, "avg_logprob": -1.2080620430611275, "compression_ratio": 1.2697841726618706, "no_speech_prob": 7.549700881820387e-13 }, { "id": 40, "seek": 12274, "start": 133.96, "end": 135.14, "text": "他说了一篇,", "tokens": [ 50926, 5000, 8090, 2289, 2257, 20878, 229, 11, 50985 ], "temperature": 0.0, "avg_logprob": -1.2080620430611275, "compression_ratio": 1.2697841726618706, "no_speech_prob": 7.549700881820387e-13 }, { "id": 41, "seek": 12274, "start": 135.14, "end": 137.92, "text": "他说了一篇,", "tokens": [ 50985, 5000, 8090, 2289, 2257, 20878, 229, 11, 51124 ], "temperature": 0.0, "avg_logprob": -1.2080620430611275, "compression_ratio": 1.2697841726618706, "no_speech_prob": 7.549700881820387e-13 }, { "id": 42, "seek": 12274, "start": 137.92, "end": 140.06, "text": "我们的认识,", "tokens": [ 51124, 15003, 1546, 7422, 97, 5233, 228, 11, 51231 ], "temperature": 0.0, "avg_logprob": -1.2080620430611275, "compression_ratio": 1.2697841726618706, "no_speech_prob": 7.549700881820387e-13 }, { "id": 43, "seek": 12274, "start": 140.06, "end": 142.06, "text": "其实是一篇的答案", "tokens": [ 51231, 41646, 1541, 2257, 20878, 229, 1546, 28223, 28899, 51331 ], "temperature": 0.0, "avg_logprob": -1.2080620430611275, "compression_ratio": 1.2697841726618706, "no_speech_prob": 7.549700881820387e-13 }, { "id": 44, "seek": 12274, "start": 142.06, "end": 143.85999999999999, "text": "合适的答案", "tokens": [ 51331, 14245, 2215, 224, 1546, 28223, 28899, 51421 ], "temperature": 0.0, "avg_logprob": -1.2080620430611275, "compression_ratio": 1.2697841726618706, "no_speech_prob": 7.549700881820387e-13 }, { "id": 45, "seek": 12274, "start": 143.85999999999999, "end": 145.82, "text": "不太不太确定,", "tokens": [ 51421, 1960, 9455, 1960, 9455, 38114, 106, 12088, 11, 51519 ], "temperature": 0.0, "avg_logprob": -1.2080620430611275, "compression_ratio": 1.2697841726618706, "no_speech_prob": 7.549700881820387e-13 }, { "id": 46, "seek": 12274, "start": 145.82, "end": 147.06, "text": "我们喜欢他们,", "tokens": [ 51519, 15003, 41606, 47911, 11, 51581 ], "temperature": 0.0, "avg_logprob": -1.2080620430611275, "compression_ratio": 1.2697841726618706, "no_speech_prob": 7.549700881820387e-13 }, { "id": 47, "seek": 12274, "start": 147.06, "end": 147.64, "text": "对吗?", "tokens": [ 51581, 8713, 14769, 30, 51610 ], "temperature": 0.0, "avg_logprob": -1.2080620430611275, "compression_ratio": 1.2697841726618706, "no_speech_prob": 7.549700881820387e-13 }, { "id": 48, "seek": 12274, "start": 147.64, "end": 148.54, "text": "例如,", "tokens": [ 51610, 17797, 8238, 11, 51655 ], "temperature": 0.0, "avg_logprob": -1.2080620430611275, "compression_ratio": 1.2697841726618706, "no_speech_prob": 7.549700881820387e-13 }, { "id": 49, "seek": 12274, "start": 148.54, "end": 150.2, "text": "他们可能会更多 empathetic", "tokens": [ 51655, 47911, 16657, 12949, 19002, 6392, 27155, 3532, 51738 ], "temperature": 0.0, "avg_logprob": -1.2080620430611275, "compression_ratio": 1.2697841726618706, "no_speech_prob": 7.549700881820387e-13 }, { "id": 50, "seek": 12274, "start": 150.2, "end": 152.1, "text": "或是更多理解我们的", "tokens": [ 51738, 19780, 1541, 19002, 6392, 13876, 17278, 15003, 1546, 51833 ], "temperature": 0.0, "avg_logprob": -1.2080620430611275, "compression_ratio": 1.2697841726618706, "no_speech_prob": 7.549700881820387e-13 }, { "id": 51, "seek": 15210, "start": 152.1, "end": 155.29999999999998, "text": "但是,他們的問題是他們的解答問題", "tokens": [ 50365, 11189, 11, 20486, 1546, 17197, 1541, 20486, 1546, 17278, 28223, 17197, 50525 ], "temperature": 0.0, "avg_logprob": -1.0324538548787434, "compression_ratio": 1.4358108108108107, "no_speech_prob": 7.037680442752159e-13 }, { "id": 52, "seek": 15210, "start": 155.29999999999998, "end": 156.79999999999998, "text": "人們不可能會不喜歡,", "tokens": [ 50525, 4035, 4623, 1960, 16657, 6236, 1960, 21385, 11, 50600 ], "temperature": 0.0, "avg_logprob": -1.0324538548787434, "compression_ratio": 1.4358108108108107, "no_speech_prob": 7.037680442752159e-13 }, { "id": 53, "seek": 15210, "start": 156.79999999999998, "end": 159.7, "text": "甚至是有個方法的方法。", "tokens": [ 50600, 17556, 20844, 1541, 2412, 3338, 9249, 11148, 1546, 9249, 11148, 1543, 50745 ], "temperature": 0.0, "avg_logprob": -1.0324538548787434, "compression_ratio": 1.4358108108108107, "no_speech_prob": 7.037680442752159e-13 }, { "id": 54, "seek": 15210, "start": 159.7, "end": 164.1, "text": "這帶來更加盡量的信息和 consistency。", "tokens": [ 50745, 2664, 32379, 3763, 19002, 9990, 5419, 94, 26748, 1546, 17665, 26460, 12565, 14416, 1543, 50965 ], "temperature": 0.0, "avg_logprob": -1.0324538548787434, "compression_ratio": 1.4358108108108107, "no_speech_prob": 7.037680442752159e-13 }, { "id": 55, "seek": 15210, "start": 164.1, "end": 167.4, "text": "但是,他們的解答案是叫做RLCD,", "tokens": [ 50965, 11189, 11, 20486, 1546, 17278, 28223, 28899, 1541, 19855, 10907, 10740, 16508, 11, 51130 ], "temperature": 0.0, "avg_logprob": -1.0324538548787434, "compression_ratio": 1.4358108108108107, "no_speech_prob": 7.037680442752159e-13 }, { "id": 56, "seek": 15210, "start": 167.4, "end": 170.79999999999998, "text": "名為 Reinforcement Learning from Calibrated Decisions。", "tokens": [ 51130, 15940, 6344, 42116, 9382, 15205, 490, 3511, 897, 5468, 12427, 4252, 1543, 51300 ], "temperature": 0.0, "avg_logprob": -1.0324538548787434, "compression_ratio": 1.4358108108108107, "no_speech_prob": 7.037680442752159e-13 }, { "id": 57, "seek": 15210, "start": 170.79999999999998, "end": 172.79999999999998, "text": "它指的是,在 Reinforcement Learning,", "tokens": [ 51300, 11284, 25922, 24620, 11, 3581, 42116, 9382, 15205, 11, 51400 ], "temperature": 0.0, "avg_logprob": -1.0324538548787434, "compression_ratio": 1.4358108108108107, "no_speech_prob": 7.037680442752159e-13 }, { "id": 58, "seek": 15210, "start": 172.79999999999998, "end": 174.6, "text": "我實在是在調整的解答案是", "tokens": [ 51400, 1654, 10376, 3581, 1541, 3581, 32943, 27662, 1546, 17278, 28223, 28899, 1541, 51490 ], "temperature": 0.0, "avg_logprob": -1.0324538548787434, "compression_ratio": 1.4358108108108107, "no_speech_prob": 7.037680442752159e-13 }, { "id": 59, "seek": 15210, "start": 174.6, "end": 176.6, "text": "的解答案是一個解答案。", "tokens": [ 51490, 1546, 17278, 28223, 28899, 1541, 8990, 17278, 28223, 28899, 1543, 51590 ], "temperature": 0.0, "avg_logprob": -1.0324538548787434, "compression_ratio": 1.4358108108108107, "no_speech_prob": 7.037680442752159e-13 }, { "id": 60, "seek": 15210, "start": 176.6, "end": 179.79999999999998, "text": "解答案是每次它提供答案,", "tokens": [ 51590, 17278, 28223, 28899, 1541, 23664, 9487, 11284, 20949, 3254, 249, 28223, 28899, 11, 51750 ], "temperature": 0.0, "avg_logprob": -1.0324538548787434, "compression_ratio": 1.4358108108108107, "no_speech_prob": 7.037680442752159e-13 }, { "id": 61, "seek": 15210, "start": 179.79999999999998, "end": 181.6, "text": "它提供一個 probability。", "tokens": [ 51750, 11284, 20949, 3254, 249, 8990, 8482, 1543, 51840 ], "temperature": 0.0, "avg_logprob": -1.0324538548787434, "compression_ratio": 1.4358108108108107, "no_speech_prob": 7.037680442752159e-13 }, { "id": 62, "seek": 18160, "start": 181.6, "end": 182.76, "text": "A confidence score", "tokens": [ 50365, 32, 6687, 6175, 50423 ], "temperature": 0.0, "avg_logprob": -0.19867820739746095, "compression_ratio": 1.6204379562043796, "no_speech_prob": 3.05654351655249e-12 }, { "id": 63, "seek": 18160, "start": 182.76, "end": 185.16, "text": "For instance, with this billing inquiry", "tokens": [ 50423, 12587, 5197, 11, 365, 341, 35618, 25736, 50543 ], "temperature": 0.0, "avg_logprob": -0.19867820739746095, "compression_ratio": 1.6204379562043796, "no_speech_prob": 3.05654351655249e-12 }, { "id": 64, "seek": 18160, "start": 185.16, "end": 189.76, "text": "the probability that I should route it to the billing department is about 90%", "tokens": [ 50543, 3322, 8482, 300, 286, 820, 7955, 309, 281, 264, 35618, 5882, 307, 466, 4289, 4, 50773 ], "temperature": 0.0, "avg_logprob": -0.19867820739746095, "compression_ratio": 1.6204379562043796, "no_speech_prob": 3.05654351655249e-12 }, { "id": 65, "seek": 18160, "start": 189.76, "end": 192.35999999999999, "text": "But if the customer also mentioned other issues", "tokens": [ 50773, 7835, 498, 264, 5474, 611, 2835, 661, 2663, 50903 ], "temperature": 0.0, "avg_logprob": -0.19867820739746095, "compression_ratio": 1.6204379562043796, "no_speech_prob": 3.05654351655249e-12 }, { "id": 66, "seek": 18160, "start": 192.35999999999999, "end": 196.35999999999999, "text": "I believe there's a certain probability it should be prioritized for other departments", "tokens": [ 50903, 40, 1697, 456, 311, 257, 1629, 8482, 309, 820, 312, 14846, 1602, 337, 661, 15326, 51103 ], "temperature": 0.0, "avg_logprob": -0.19867820739746095, "compression_ratio": 1.6204379562043796, "no_speech_prob": 3.05654351655249e-12 }, { "id": 67, "seek": 18160, "start": 196.35999999999999, "end": 200.16, "text": "So I might assign a 10% or 20% probability to those as well", "tokens": [ 51103, 6455, 286, 1062, 6269, 257, 1266, 4, 420, 945, 4, 8482, 281, 729, 382, 731, 51293 ], "temperature": 0.0, "avg_logprob": -0.19867820739746095, "compression_ratio": 1.6204379562043796, "no_speech_prob": 3.05654351655249e-12 }, { "id": 68, "seek": 18160, "start": 200.16, "end": 202.95999999999998, "text": "At this point, you can set a threshold", "tokens": [ 51293, 18684, 341, 935, 11, 291, 393, 992, 257, 14678, 51433 ], "temperature": 0.0, "avg_logprob": -0.19867820739746095, "compression_ratio": 1.6204379562043796, "no_speech_prob": 3.05654351655249e-12 }, { "id": 69, "seek": 18160, "start": 202.95999999999998, "end": 206.35999999999999, "text": "For example, if the churn risk is very high", "tokens": [ 51433, 12587, 1365, 11, 498, 264, 417, 925, 3148, 307, 588, 1090, 51603 ], "temperature": 0.0, "avg_logprob": -0.19867820739746095, "compression_ratio": 1.6204379562043796, "no_speech_prob": 3.05654351655249e-12 }, { "id": 70, "seek": 18160, "start": 206.35999999999999, "end": 208.76, "text": "I can have a human intervene directly", "tokens": [ 51603, 40, 393, 362, 257, 1952, 30407, 3838, 51723 ], "temperature": 0.0, "avg_logprob": -0.19867820739746095, "compression_ratio": 1.6204379562043796, "no_speech_prob": 3.05654351655249e-12 }, { "id": 71, "seek": 20876, "start": 208.76, "end": 212.76, "text": "如果它是低,我可以让它 AI model handle the response.", "tokens": [ 50365, 13119, 11284, 1541, 41377, 11, 1654, 6723, 33650, 11284, 7318, 2316, 4813, 264, 4134, 13, 50565 ], "temperature": 0.0, "avg_logprob": -1.17284916469029, "compression_ratio": 1.2976588628762542, "no_speech_prob": 1.2993192676294396e-12 }, { "id": 72, "seek": 20876, "start": 212.76, "end": 216.76, "text": "所以,它帮助我们去解决哪个性能量的权量使用", "tokens": [ 50565, 7239, 11, 11284, 4845, 106, 37618, 15003, 6734, 17278, 5676, 111, 17028, 7549, 21686, 8225, 26748, 1546, 4422, 225, 26748, 22982, 9254, 50765 ], "temperature": 0.0, "avg_logprob": -1.17284916469029, "compression_ratio": 1.2976588628762542, "no_speech_prob": 1.2993192676294396e-12 }, { "id": 73, "seek": 20876, "start": 216.76, "end": 219.51999999999998, "text": "来解决于不同的情况下。", "tokens": [ 50765, 6912, 17278, 5676, 111, 37732, 47123, 1546, 46514, 4438, 1543, 50903 ], "temperature": 0.0, "avg_logprob": -1.17284916469029, "compression_ratio": 1.2976588628762542, "no_speech_prob": 1.2993192676294396e-12 }, { "id": 74, "seek": 20876, "start": 219.51999999999998, "end": 224.23999999999998, "text": "所以,我相信这个模型会将将其进行我们的生活生活", "tokens": [ 50903, 7239, 11, 1654, 35428, 15368, 41908, 39823, 12949, 45456, 45456, 9572, 36700, 8082, 15003, 1546, 49958, 49958, 51139 ], "temperature": 0.0, "avg_logprob": -1.17284916469029, "compression_ratio": 1.2976588628762542, "no_speech_prob": 1.2993192676294396e-12 }, { "id": 75, "seek": 20876, "start": 224.23999999999998, "end": 226.35999999999999, "text": "和很多不同的专业。", "tokens": [ 51139, 12565, 20778, 47123, 1546, 940, 241, 940, 248, 1543, 51245 ], "temperature": 0.0, "avg_logprob": -1.17284916469029, "compression_ratio": 1.2976588628762542, "no_speech_prob": 1.2993192676294396e-12 }, { "id": 76, "seek": 20876, "start": 226.35999999999999, "end": 230.2, "text": "以前我们曾经用了一枪创造的创造,", "tokens": [ 51245, 44996, 15003, 43801, 30276, 9254, 2289, 2257, 7360, 103, 2437, 249, 37583, 1546, 2437, 249, 37583, 11, 51437 ], "temperature": 0.0, "avg_logprob": -1.17284916469029, "compression_ratio": 1.2976588628762542, "no_speech_prob": 1.2993192676294396e-12 }, { "id": 77, "seek": 20876, "start": 230.2, "end": 233.0, "text": "或者我们不需要太多的时间。", "tokens": [ 51437, 31148, 15003, 1960, 35748, 9455, 6392, 1546, 44848, 1543, 51577 ], "temperature": 0.0, "avg_logprob": -1.17284916469029, "compression_ratio": 1.2976588628762542, "no_speech_prob": 1.2993192676294396e-12 }, { "id": 78, "seek": 20876, "start": 233.0, "end": 237.48, "text": "但这些随意的随意的选择是我们的选择。", "tokens": [ 51577, 8395, 5562, 13824, 10673, 237, 9042, 1546, 10673, 237, 9042, 1546, 2215, 231, 6852, 102, 1541, 15003, 1546, 2215, 231, 6852, 102, 1543, 51801 ], "temperature": 0.0, "avg_logprob": -1.17284916469029, "compression_ratio": 1.2976588628762542, "no_speech_prob": 1.2993192676294396e-12 }, { "id": 79, "seek": 23748, "start": 237.48, "end": 242.6, "text": "但是在这个新的模式中,我们的选择和认证能更加更加更加适合。", "tokens": [ 50365, 11189, 3581, 15368, 12560, 1546, 41908, 27584, 5975, 11, 15003, 1546, 2215, 231, 6852, 102, 12565, 7422, 97, 5233, 223, 8225, 19002, 9990, 19002, 9990, 19002, 9990, 2215, 224, 14245, 1543, 50621 ], "temperature": 0.0, "avg_logprob": -1.115296698235846, "compression_ratio": 1.578740157480315, "no_speech_prob": 1.1378603137837673e-12 }, { "id": 80, "seek": 23748, "start": 242.6, "end": 246.67999999999998, "text": "如果我们说我们的选择这个进入我们的页务,", "tokens": [ 50621, 13119, 15003, 8090, 15003, 1546, 2215, 231, 6852, 102, 15368, 36700, 14028, 15003, 1546, 10178, 113, 5087, 94, 11, 50825 ], "temperature": 0.0, "avg_logprob": -1.115296698235846, "compression_ratio": 1.578740157480315, "no_speech_prob": 1.1378603137837673e-12 }, { "id": 81, "seek": 23748, "start": 246.67999999999998, "end": 251.88, "text": "我们肯定能够用这个模式,因为现在,我们的页务员工程,", "tokens": [ 50825, 15003, 14356, 107, 12088, 8225, 1787, 253, 9254, 15368, 41908, 27584, 11, 34627, 25040, 11, 15003, 1546, 10178, 113, 5087, 94, 3606, 246, 23323, 29649, 11, 51085 ], "temperature": 0.0, "avg_logprob": -1.115296698235846, "compression_ratio": 1.578740157480315, "no_speech_prob": 1.1378603137837673e-12 }, { "id": 82, "seek": 23748, "start": 251.88, "end": 255.64, "text": "他们其实会有很多认证和评价的评价。", "tokens": [ 51085, 47911, 41646, 12949, 2412, 20778, 7422, 97, 5233, 223, 12565, 5233, 226, 1550, 115, 1546, 5233, 226, 1550, 115, 1543, 51273 ], "temperature": 0.0, "avg_logprob": -1.115296698235846, "compression_ratio": 1.578740157480315, "no_speech_prob": 1.1378603137837673e-12 }, { "id": 83, "seek": 23748, "start": 255.64, "end": 259.48, "text": "例如,如果我们的收入的收入我们的预期,", "tokens": [ 51273, 17797, 8238, 11, 13119, 15003, 1546, 18681, 14028, 1546, 18681, 14028, 15003, 1546, 12501, 226, 16786, 11, 51465 ], "temperature": 0.0, "avg_logprob": -1.115296698235846, "compression_ratio": 1.578740157480315, "no_speech_prob": 1.1378603137837673e-12 }, { "id": 84, "seek": 23748, "start": 259.48, "end": 261.71999999999997, "text": "我们还是需要这些认证,", "tokens": [ 51465, 15003, 45726, 35748, 5562, 13824, 7422, 97, 5233, 223, 11, 51577 ], "temperature": 0.0, "avg_logprob": -1.115296698235846, "compression_ratio": 1.578740157480315, "no_speech_prob": 1.1378603137837673e-12 }, { "id": 85, "seek": 23748, "start": 261.71999999999997, "end": 264.68, "text": "这会常常需要一个人的路程式的方法。", "tokens": [ 51577, 5562, 12949, 11279, 11279, 35748, 20182, 4035, 1546, 24658, 29649, 27584, 1546, 9249, 11148, 1543, 51725 ], "temperature": 0.0, "avg_logprob": -1.115296698235846, "compression_ratio": 1.578740157480315, "no_speech_prob": 1.1378603137837673e-12 }, { "id": 86, "seek": 26468, "start": 264.68, "end": 266.6, "text": "但是如果我们有一个模式的模式,", "tokens": [ 50365, 11189, 13119, 15003, 2412, 20182, 41908, 27584, 1546, 41908, 27584, 11, 50461 ], "temperature": 0.0, "avg_logprob": -1.1311071644658628, "compression_ratio": 1.5344827586206897, "no_speech_prob": 7.943947149396946e-13 }, { "id": 87, "seek": 26468, "start": 266.6, "end": 269.8, "text": "我们可以定义它的质量从开始", "tokens": [ 50461, 15003, 6723, 12088, 2930, 231, 45224, 18464, 101, 26748, 35630, 45213, 50621 ], "temperature": 0.0, "avg_logprob": -1.1311071644658628, "compression_ratio": 1.5344827586206897, "no_speech_prob": 7.943947149396946e-13 }, { "id": 88, "seek": 26468, "start": 269.8, "end": 274.08, "text": "即使是80%,60%,40%的距离,", "tokens": [ 50621, 39127, 22982, 1541, 4702, 8923, 4550, 8923, 5254, 4, 1546, 6563, 251, 18614, 119, 11, 50835 ], "temperature": 0.0, "avg_logprob": -1.1311071644658628, "compression_ratio": 1.5344827586206897, "no_speech_prob": 7.943947149396946e-13 }, { "id": 89, "seek": 26468, "start": 274.08, "end": 278.52, "text": "我们只需要在一个很低的40%的距离。", "tokens": [ 50835, 15003, 14003, 35748, 3581, 20182, 4563, 41377, 1546, 5254, 4, 1546, 6563, 251, 18614, 119, 1543, 51057 ], "temperature": 0.0, "avg_logprob": -1.1311071644658628, "compression_ratio": 1.5344827586206897, "no_speech_prob": 7.943947149396946e-13 }, { "id": 90, "seek": 26468, "start": 278.52, "end": 281.08, "text": "如果它基本上是在结构的质量,", "tokens": [ 51057, 13119, 11284, 37946, 5708, 1541, 3581, 45641, 7360, 226, 1546, 18464, 101, 26748, 11, 51185 ], "temperature": 0.0, "avg_logprob": -1.1311071644658628, "compression_ratio": 1.5344827586206897, "no_speech_prob": 7.943947149396946e-13 }, { "id": 91, "seek": 26468, "start": 281.08, "end": 283.48, "text": "我们不需要保持人的质量,", "tokens": [ 51185, 15003, 1960, 35748, 24302, 17694, 4035, 1546, 18464, 101, 26748, 11, 51305 ], "temperature": 0.0, "avg_logprob": -1.1311071644658628, "compression_ratio": 1.5344827586206897, "no_speech_prob": 7.943947149396946e-13 }, { "id": 92, "seek": 26468, "start": 283.48, "end": 286.6, "text": "我们不需要保持人的质量。", "tokens": [ 51305, 15003, 1960, 35748, 24302, 17694, 4035, 1546, 18464, 101, 26748, 1543, 51461 ], "temperature": 0.0, "avg_logprob": -1.1311071644658628, "compression_ratio": 1.5344827586206897, "no_speech_prob": 7.943947149396946e-13 }, { "id": 93, "seek": 26468, "start": 286.6, "end": 290.04, "text": "所以,在这个将来的更多的模式,", "tokens": [ 51461, 7239, 11, 3581, 15368, 45456, 6912, 1546, 19002, 6392, 1546, 41908, 27584, 11, 51633 ], "temperature": 0.0, "avg_logprob": -1.1311071644658628, "compression_ratio": 1.5344827586206897, "no_speech_prob": 7.943947149396946e-13 }, { "id": 94, "seek": 26468, "start": 290.04, "end": 292.28000000000003, "text": "它可以更加大幅并且在这些模式上", "tokens": [ 51633, 11284, 6723, 19002, 9990, 3582, 3509, 227, 3509, 114, 20334, 3581, 5562, 13824, 41908, 27584, 5708, 51745 ], "temperature": 0.0, "avg_logprob": -1.1311071644658628, "compression_ratio": 1.5344827586206897, "no_speech_prob": 7.943947149396946e-13 }, { "id": 95, "seek": 29228, "start": 292.28, "end": 295.28, "text": "这个模式其实只发了一个新版本", "tokens": [ 50365, 15368, 41908, 27584, 41646, 14003, 28926, 2289, 20182, 12560, 42096, 8802, 50515 ], "temperature": 0.0, "avg_logprob": -1.0488124593098958, "compression_ratio": 1.2867924528301886, "no_speech_prob": 1.0198772165334602e-12 }, { "id": 96, "seek": 29228, "start": 295.28, "end": 299.03999999999996, "text": "没有提供的技术和计程度", "tokens": [ 50515, 17944, 20949, 3254, 249, 1546, 32502, 1474, 107, 12565, 7422, 94, 29649, 13127, 50703 ], "temperature": 0.0, "avg_logprob": -1.0488124593098958, "compression_ratio": 1.2867924528301886, "no_speech_prob": 1.0198772165334602e-12 }, { "id": 97, "seek": 29228, "start": 299.03999999999996, "end": 301.96, "text": "我觉得这个方法非常值得考虑", "tokens": [ 50703, 45681, 15368, 9249, 11148, 14392, 40242, 5916, 26504, 12026, 239, 50849 ], "temperature": 0.0, "avg_logprob": -1.0488124593098958, "compression_ratio": 1.2867924528301886, "no_speech_prob": 1.0198772165334602e-12 }, { "id": 98, "seek": 29228, "start": 301.96, "end": 304.15999999999997, "text": "它使用一个系统一式的方法", "tokens": [ 50849, 11284, 22982, 9254, 20182, 25368, 10115, 253, 2257, 27584, 1546, 9249, 11148, 50959 ], "temperature": 0.0, "avg_logprob": -1.0488124593098958, "compression_ratio": 1.2867924528301886, "no_speech_prob": 1.0198772165334602e-12 }, { "id": 99, "seek": 29228, "start": 304.15999999999997, "end": 305.64, "text": "是快思考", "tokens": [ 50959, 1541, 10251, 8870, 26504, 51033 ], "temperature": 0.0, "avg_logprob": -1.0488124593098958, "compression_ratio": 1.2867924528301886, "no_speech_prob": 1.0198772165334602e-12 }, { "id": 100, "seek": 29228, "start": 305.64, "end": 306.88, "text": " Daniel Kahneman的作词", "tokens": [ 51033, 8033, 591, 12140, 15023, 1546, 11914, 5233, 235, 51095 ], "temperature": 0.0, "avg_logprob": -1.0488124593098958, "compression_ratio": 1.2867924528301886, "no_speech_prob": 1.0198772165334602e-12 }, { "id": 101, "seek": 29228, "start": 306.88, "end": 309.28, "text": "说一大部分人的认识", "tokens": [ 51095, 8090, 2257, 3582, 32174, 4035, 1546, 7422, 97, 5233, 228, 51215 ], "temperature": 0.0, "avg_logprob": -1.0488124593098958, "compression_ratio": 1.2867924528301886, "no_speech_prob": 1.0198772165334602e-12 }, { "id": 102, "seek": 29228, "start": 309.28, "end": 311.35999999999996, "text": "不在实际上的认识", "tokens": [ 51215, 1960, 3581, 24726, 8842, 227, 5708, 1546, 7422, 97, 5233, 228, 51319 ], "temperature": 0.0, "avg_logprob": -1.0488124593098958, "compression_ratio": 1.2867924528301886, "no_speech_prob": 1.0198772165334602e-12 }, { "id": 103, "seek": 29228, "start": 311.35999999999996, "end": 312.71999999999997, "text": "我们立刻的认识", "tokens": [ 51319, 15003, 24409, 45500, 1546, 7422, 97, 5233, 228, 51387 ], "temperature": 0.0, "avg_logprob": -1.0488124593098958, "compression_ratio": 1.2867924528301886, "no_speech_prob": 1.0198772165334602e-12 }, { "id": 104, "seek": 29228, "start": 312.71999999999997, "end": 314.64, "text": "我们立刻的认识", "tokens": [ 51387, 15003, 24409, 45500, 1546, 7422, 97, 5233, 228, 51483 ], "temperature": 0.0, "avg_logprob": -1.0488124593098958, "compression_ratio": 1.2867924528301886, "no_speech_prob": 1.0198772165334602e-12 }, { "id": 105, "seek": 29228, "start": 314.64, "end": 317.44, "text": "例如,选择哪个衣服", "tokens": [ 51483, 17797, 8238, 11, 2215, 231, 6852, 102, 17028, 7549, 9890, 96, 27408, 51623 ], "temperature": 0.0, "avg_logprob": -1.0488124593098958, "compression_ratio": 1.2867924528301886, "no_speech_prob": 1.0198772165334602e-12 }, { "id": 106, "seek": 29228, "start": 317.44, "end": 318.88, "text": "很自然的衣服", "tokens": [ 51623, 4563, 9722, 5823, 1546, 9890, 96, 27408, 51695 ], "temperature": 0.0, "avg_logprob": -1.0488124593098958, "compression_ratio": 1.2867924528301886, "no_speech_prob": 1.0198772165334602e-12 }, { "id": 107, "seek": 29228, "start": 318.88, "end": 321.44, "text": "我只穿着衣服", "tokens": [ 51695, 1654, 14003, 13740, 123, 20708, 9890, 96, 27408, 51823 ], "temperature": 0.0, "avg_logprob": -1.0488124593098958, "compression_ratio": 1.2867924528301886, "no_speech_prob": 1.0198772165334602e-12 }, { "id": 108, "seek": 32144, "start": 321.44, "end": 324.2, "text": "and when I'm going to work I might wear a different pair.", "tokens": [ 50365, 474, 562, 286, 478, 516, 281, 589, 286, 1062, 3728, 257, 819, 6119, 13, 50503 ], "temperature": 0.0, "avg_logprob": -0.1788664271688869, "compression_ratio": 1.6347517730496455, "no_speech_prob": 7.375773169310262e-13 }, { "id": 109, "seek": 32144, "start": 324.2, "end": 326.94, "text": "So this judgment is actually very intuitive.", "tokens": [ 50503, 6455, 341, 12216, 307, 767, 588, 21769, 13, 50640 ], "temperature": 0.0, "avg_logprob": -0.1788664271688869, "compression_ratio": 1.6347517730496455, "no_speech_prob": 7.375773169310262e-13 }, { "id": 110, "seek": 32144, "start": 326.94, "end": 329.28, "text": "It's the same for AI models.", "tokens": [ 50640, 3522, 311, 264, 912, 337, 7318, 5245, 13, 50757 ], "temperature": 0.0, "avg_logprob": -0.1788664271688869, "compression_ratio": 1.6347517730496455, "no_speech_prob": 7.375773169310262e-13 }, { "id": 111, "seek": 32144, "start": 329.28, "end": 334.04, "text": "You don't actually need to generate or think through every single judgment", "tokens": [ 50757, 3223, 500, 380, 767, 643, 281, 8460, 420, 519, 807, 633, 2167, 12216, 50995 ], "temperature": 0.0, "avg_logprob": -0.1788664271688869, "compression_ratio": 1.6347517730496455, "no_speech_prob": 7.375773169310262e-13 }, { "id": 112, "seek": 32144, "start": 334.04, "end": 336.2, "text": "or use chain of thought reasoning.", "tokens": [ 50995, 284, 764, 5021, 295, 1194, 21577, 13, 51103 ], "temperature": 0.0, "avg_logprob": -0.1788664271688869, "compression_ratio": 1.6347517730496455, "no_speech_prob": 7.375773169310262e-13 }, { "id": 113, "seek": 32144, "start": 336.2, "end": 340.2, "text": "I can use a smaller model or a model designed for fast thinking", "tokens": [ 51103, 40, 393, 764, 257, 4356, 2316, 420, 257, 2316, 4761, 337, 2370, 1953, 51303 ], "temperature": 0.0, "avg_logprob": -0.1788664271688869, "compression_ratio": 1.6347517730496455, "no_speech_prob": 7.375773169310262e-13 }, { "id": 114, "seek": 32144, "start": 340.2, "end": 341.68, "text": "with parallel computing", "tokens": [ 51303, 11820, 8952, 15866, 51377 ], "temperature": 0.0, "avg_logprob": -0.1788664271688869, "compression_ratio": 1.6347517730496455, "no_speech_prob": 7.375773169310262e-13 }, { "id": 115, "seek": 32144, "start": 341.68, "end": 344.68, "text": "to accomplish our tasks and goals very efficiently.", "tokens": [ 51377, 1353, 9021, 527, 9608, 293, 5493, 588, 19621, 13, 51527 ], "temperature": 0.0, "avg_logprob": -0.1788664271688869, "compression_ratio": 1.6347517730496455, "no_speech_prob": 7.375773169310262e-13 }, { "id": 116, "seek": 32144, "start": 344.68, "end": 348.6, "text": "So I think it might really make its way into our society soon", "tokens": [ 51527, 6455, 286, 519, 309, 1062, 534, 652, 1080, 636, 666, 527, 4086, 2321, 51723 ], "temperature": 0.0, "avg_logprob": -0.1788664271688869, "compression_ratio": 1.6347517730496455, "no_speech_prob": 7.375773169310262e-13 }, { "id": 117, "seek": 32144, "start": 348.6, "end": 350.7, "text": "because it's cheap, right?", "tokens": [ 51723, 17566, 309, 311, 7084, 11, 558, 30, 51828 ], "temperature": 0.0, "avg_logprob": -0.1788664271688869, "compression_ratio": 1.6347517730496455, "no_speech_prob": 7.375773169310262e-13 }, { "id": 118, "seek": 35070, "start": 350.7, "end": 353.78, "text": "而且是最高的,因為它是很高的,", "tokens": [ 50365, 22942, 1541, 8661, 12979, 1546, 11, 11471, 11284, 1541, 4563, 12979, 1546, 11, 50519 ], "temperature": 0.0, "avg_logprob": -0.6109539794921875, "compression_ratio": 2.537254901960784, "no_speech_prob": 1.7009253163932647e-12 }, { "id": 119, "seek": 35070, "start": 353.78, "end": 357.09999999999997, "text": "在0.042萬元的價值,", "tokens": [ 50519, 3581, 15, 13, 14565, 17, 14099, 14812, 1546, 45441, 40242, 11, 50685 ], "temperature": 0.0, "avg_logprob": -0.6109539794921875, "compression_ratio": 2.537254901960784, "no_speech_prob": 1.7009253163932647e-12 }, { "id": 120, "seek": 35070, "start": 357.09999999999997, "end": 359.78, "text": "比 GPT-4的價值,可能是10萬元。", "tokens": [ 50685, 11706, 26039, 51, 12, 19, 1546, 45441, 40242, 11, 16657, 1541, 3279, 14099, 14812, 1543, 50819 ], "temperature": 0.0, "avg_logprob": -0.6109539794921875, "compression_ratio": 2.537254901960784, "no_speech_prob": 1.7009253163932647e-12 }, { "id": 121, "seek": 35070, "start": 359.78, "end": 363.3, "text": "價值是這麼高,他們現在不在賣現在的價值。", "tokens": [ 50819, 45441, 40242, 1541, 21269, 12979, 11, 20486, 12648, 1960, 3581, 13352, 96, 12648, 1546, 45441, 40242, 1543, 50995 ], "temperature": 0.0, "avg_logprob": -0.6109539794921875, "compression_ratio": 2.537254901960784, "no_speech_prob": 1.7009253163932647e-12 }, { "id": 122, "seek": 35070, "start": 363.3, "end": 367.82, "text": "所以它可以使用在一大多數的評測 scenarios。", "tokens": [ 50995, 7239, 11284, 6723, 22982, 9254, 3581, 2257, 3582, 6392, 30622, 1546, 39115, 9592, 105, 15077, 1543, 51221 ], "temperature": 0.0, "avg_logprob": -0.6109539794921875, "compression_ratio": 2.537254901960784, "no_speech_prob": 1.7009253163932647e-12 }, { "id": 123, "seek": 35070, "start": 367.82, "end": 370.74, "text": "我們看到每個小型的手法", "tokens": [ 51221, 5884, 18032, 23664, 3338, 7322, 39823, 1546, 11389, 11148, 51367 ], "temperature": 0.0, "avg_logprob": -0.6109539794921875, "compression_ratio": 2.537254901960784, "no_speech_prob": 1.7009253163932647e-12 }, { "id": 124, "seek": 35070, "start": 370.74, "end": 373.41999999999996, "text": "可能有一個決定的模式背後。", "tokens": [ 51367, 16657, 2412, 8990, 33540, 12088, 1546, 41908, 27584, 46329, 5661, 1543, 51501 ], "temperature": 0.0, "avg_logprob": -0.6109539794921875, "compression_ratio": 2.537254901960784, "no_speech_prob": 1.7009253163932647e-12 }, { "id": 125, "seek": 37342, "start": 373.42, "end": 378.58000000000004, "text": "以前,一名 classifier would have required someone to spend ages training it, right?", "tokens": [ 50365, 44996, 11, 2257, 15940, 1508, 9902, 576, 362, 4739, 1580, 281, 3496, 12357, 3097, 309, 11, 558, 30, 50623 ], "temperature": 0.0, "avg_logprob": -1.1673861416903408, "compression_ratio": 1.193220338983051, "no_speech_prob": 1.3510860182378015e-12 }, { "id": 126, "seek": 37342, "start": 378.58000000000004, "end": 382.74, "text": "我們以前要有的要求,但現在是沒有必要的", "tokens": [ 50623, 5884, 44996, 4275, 2412, 1546, 4275, 32718, 11, 8395, 12648, 1541, 6963, 28531, 4275, 1546, 50831 ], "temperature": 0.0, "avg_logprob": -1.1673861416903408, "compression_ratio": 1.193220338983051, "no_speech_prob": 1.3510860182378015e-12 }, { "id": 127, "seek": 37342, "start": 382.74, "end": 385.78000000000003, "text": "現在,這 model可以幫我們做了一切的評判,", "tokens": [ 50831, 12648, 11, 2664, 2316, 6723, 32187, 5884, 10907, 2289, 2257, 23632, 1546, 39115, 2437, 97, 11, 50983 ], "temperature": 0.0, "avg_logprob": -1.1673861416903408, "compression_ratio": 1.193220338983051, "no_speech_prob": 1.3510860182378015e-12 }, { "id": 128, "seek": 37342, "start": 385.78000000000003, "end": 388.66, "text": "決定定 specific outcomes based on the context", "tokens": [ 50983, 33540, 12088, 12088, 2685, 10070, 2361, 322, 264, 4319, 51127 ], "temperature": 0.0, "avg_logprob": -1.1673861416903408, "compression_ratio": 1.193220338983051, "no_speech_prob": 1.3510860182378015e-12 }, { "id": 129, "seek": 37342, "start": 388.66, "end": 394.46000000000004, "text": "這代表,這代表,已經融化了更多的技術和考驗", "tokens": [ 51127, 2664, 19105, 17571, 11, 2664, 19105, 17571, 11, 25338, 43772, 235, 23756, 2289, 19002, 6392, 1546, 32502, 9890, 241, 12565, 26504, 24023, 245, 51417 ], "temperature": 0.0, "avg_logprob": -1.1673861416903408, "compression_ratio": 1.193220338983051, "no_speech_prob": 1.3510860182378015e-12 }, { "id": 130, "seek": 37342, "start": 394.46000000000004, "end": 398.86, "text": "我們會看到這個功能伸展開更多的使用", "tokens": [ 51417, 5884, 6236, 18032, 6287, 32311, 8225, 7384, 116, 43491, 8949, 19002, 6392, 1546, 22982, 9254, 51637 ], "temperature": 0.0, "avg_logprob": -1.1673861416903408, "compression_ratio": 1.193220338983051, "no_speech_prob": 1.3510860182378015e-12 }, { "id": 131, "seek": 39886, "start": 398.86, "end": 402.62, "text": "是的,我认为这个是非常有趣的新方式,", "tokens": [ 50365, 30470, 11, 1654, 7422, 97, 13992, 15368, 1541, 14392, 2412, 39835, 1546, 12560, 9249, 27584, 11, 50553 ], "temperature": 0.0, "avg_logprob": -0.2115525648328993, "compression_ratio": 3.4862385321100917, "no_speech_prob": 1.3199455634396728e-12 }, { "id": 132, "seek": 40262, "start": 402.62, "end": 406.38, "text": "一遍完全不同的方法使用力", "tokens": [ 50365, 2257, 3330, 235, 37100, 47123, 1546, 9249, 11148, 22982, 9254, 13486, 50553 ], "temperature": 0.0, "avg_logprob": -1.4061235700334822, "compression_ratio": 1.5045045045045045, "no_speech_prob": 1.2793504320166016e-12 }, { "id": 133, "seek": 40262, "start": 406.38, "end": 407.98, "text": "我們之前說明,", "tokens": [ 50553, 5884, 32442, 4622, 11100, 11, 50633 ], "temperature": 0.0, "avg_logprob": -1.4061235700334822, "compression_ratio": 1.5045045045045045, "no_speech_prob": 1.2793504320166016e-12 }, { "id": 134, "seek": 40262, "start": 407.98, "end": 410.7, "text": "未來可能會有小型的模式", "tokens": [ 50633, 29954, 3763, 16657, 6236, 2412, 7322, 39823, 1546, 41908, 27584, 50769 ], "temperature": 0.0, "avg_logprob": -1.4061235700334822, "compression_ratio": 1.5045045045045045, "no_speech_prob": 1.2793504320166016e-12 }, { "id": 135, "seek": 40262, "start": 410.7, "end": 412.42, "text": "我們只需要有小型的模式", "tokens": [ 50769, 5884, 14003, 35748, 2412, 7322, 39823, 1546, 41908, 27584, 50855 ], "temperature": 0.0, "avg_logprob": -1.4061235700334822, "compression_ratio": 1.5045045045045045, "no_speech_prob": 1.2793504320166016e-12 }, { "id": 136, "seek": 40262, "start": 412.42, "end": 414.5, "text": "提供的模式", "tokens": [ 50855, 20949, 3254, 249, 1546, 41908, 27584, 50959 ], "temperature": 0.0, "avg_logprob": -1.4061235700334822, "compression_ratio": 1.5045045045045045, "no_speech_prob": 1.2793504320166016e-12 }, { "id": 137, "seek": 40262, "start": 414.5, "end": 416.5, "text": "而是用販的模式", "tokens": [ 50959, 11070, 1541, 9254, 11561, 102, 1546, 41908, 27584, 51059 ], "temperature": 0.0, "avg_logprob": -1.4061235700334822, "compression_ratio": 1.5045045045045045, "no_speech_prob": 1.2793504320166016e-12 }, { "id": 138, "seek": 40262, "start": 416.5, "end": 417.38, "text": "每個都市場的模式", "tokens": [ 51059, 23664, 3338, 7182, 27261, 21446, 1546, 41908, 27584, 51103 ], "temperature": 0.0, "avg_logprob": -1.4061235700334822, "compression_ratio": 1.5045045045045045, "no_speech_prob": 1.2793504320166016e-12 }, { "id": 139, "seek": 40262, "start": 417.38, "end": 419.62, "text": "我們最近的書", "tokens": [ 51103, 5884, 46777, 1546, 29801, 51215 ], "temperature": 0.0, "avg_logprob": -1.4061235700334822, "compression_ratio": 1.5045045045045045, "no_speech_prob": 1.2793504320166016e-12 }, { "id": 140, "seek": 40262, "start": 419.62, "end": 420.94, "text": "我們最近的書", "tokens": [ 51215, 5884, 46777, 1546, 29801, 51281 ], "temperature": 0.0, "avg_logprob": -1.4061235700334822, "compression_ratio": 1.5045045045045045, "no_speech_prob": 1.2793504320166016e-12 }, { "id": 141, "seek": 40262, "start": 420.94, "end": 422.54, "text": "也有同一個同意", "tokens": [ 51281, 6404, 2412, 13089, 8990, 13089, 9042, 51361 ], "temperature": 0.0, "avg_logprob": -1.4061235700334822, "compression_ratio": 1.5045045045045045, "no_speech_prob": 1.2793504320166016e-12 }, { "id": 142, "seek": 40262, "start": 422.54, "end": 425.06, "text": "成為我們目標的最適合的方法", "tokens": [ 51361, 11336, 6344, 5884, 11386, 46299, 1546, 8661, 3330, 102, 14245, 1546, 9249, 11148, 51487 ], "temperature": 0.0, "avg_logprob": -1.4061235700334822, "compression_ratio": 1.5045045045045045, "no_speech_prob": 1.2793504320166016e-12 }, { "id": 143, "seek": 40262, "start": 425.06, "end": 426.5, "text": "是完全不平的", "tokens": [ 51487, 1541, 37100, 1960, 16716, 1546, 51559 ], "temperature": 0.0, "avg_logprob": -1.4061235700334822, "compression_ratio": 1.5045045045045045, "no_speech_prob": 1.2793504320166016e-12 }, { "id": 144, "seek": 40262, "start": 426.5, "end": 431.18, "text": "所以我們想分享這些 JB的模式", "tokens": [ 51559, 7239, 5884, 7093, 30855, 29869, 43019, 1546, 41908, 27584, 51793 ], "temperature": 0.0, "avg_logprob": -1.4061235700334822, "compression_ratio": 1.5045045045045045, "no_speech_prob": 1.2793504320166016e-12 }, { "id": 145, "seek": 43118, "start": 431.18, "end": 435.36, "text": "They've opened their waitlist, so if you'd like to try them out, feel free to join.", "tokens": [ 50365, 8829, 600, 5625, 641, 1699, 8264, 11, 370, 498, 291, 1116, 411, 281, 853, 552, 484, 11, 841, 1737, 281, 3917, 13, 50574 ], "temperature": 0.0, "avg_logprob": -0.23044810575597427, "compression_ratio": 1.1122448979591837, "no_speech_prob": 1.155115825039743e-12 }, { "id": 146, "seek": 43118, "start": 435.82, "end": 437.68, "text": "Have a great weekend. Bye!", "tokens": [ 50597, 25659, 257, 869, 6711, 13, 4621, 0, 50690 ], "temperature": 0.0, "avg_logprob": -0.23044810575597427, "compression_ratio": 1.1122448979591837, "no_speech_prob": 1.155115825039743e-12 } ], "language": "zh" }