تبرز الأشجار (TOT) ، وهي تقنية جديدة تستعير الأفكار من إطار الشجرة ، وتوسيع ويعزز مفهوم الطالب الفأر الراسخ ، وبالتالي تمكين النماذج اللغوية الكبيرة ، مثل ChatGPT ، لإظهار قدرات المنطق العليا. تتيح تقنية المطالبة بشجرة الفكرة هذه نماذج لغة كبيرة لتصحيح أخطائها بشكل مستقل مع تراكم المعرفة تدريجياً.
في أحد الأمثلة ، تعمل موجه TOT على تحسين قدرة التفكير في ChatGPT 3.5 على الإجابة على سؤال لا يمكن الإجابة عليه من قبل ChatGPT 4 فقط.
غالبًا ما تكافح نماذج اللغة الكبيرة (LLMS) للإجابة على أنواع معينة من الأسئلة المعقدة. إليك مثال تمثيلي:
Bob is in the living room.
He walks to the kitchen, carrying a cup.
He puts a ball in the cup and carries the cup to the bedroom.
He turns the cup upside down, then walks to the garden.
He puts the cup down in the garden, then walks to the garage.
Where is the ball?
الإجابة الصحيحة هي أن الكرة في غرفة النوم. على الرغم من أن LLMs غير محددة ، إلا أن الردود التي تلقيتها خلال عدة محاولات كانت على النحو التالي:
يوجد مفهوم معترف به على نطاق واسع المعروف باسم "سلسلة السلسلة المطلوبة" (COT for Short) ، والذي يشجع نماذج اللغة الكبيرة (LLMs) على توضيح عملية تفكيرهم ، وبالتالي زيادة احتمال الاستجابة الصحيحة. يمكن إرجاع جذور هذا المفهوم إلى ورقة من يناير 2022 بعنوان: سلسلة من الأفكار تثير التفكير في نماذج اللغة الكبيرة.
في جوهرها ، يمكن تنفيذ تقنية COT عن طريق إضافة شيء أقرب إلى ما يلي إلى موجه:
Think carefully and logically, explaining your answer.
يعزز هذا النهج أداء LLMS في الإجابة على العديد من الأسئلة ، على الرغم من اختلاف فعاليته.
إليك مثال على استجابة ChatGPT 3.5:
Based on the given sequence of events, we can logically determine the location of the ball. Let's analyze each step:
1. Bob is in the living room.
- No mention of the ball in the living room.
2. Bob walks to the kitchen, carrying a cup.
- No mention of the ball in the kitchen yet.
3. Bob puts a ball in the cup.
- The ball is now inside the cup.
4. Bob carries the cup to the bedroom.
- The ball is still inside the cup.
5. Bob turns the cup upside down.
- If the ball was inside the cup, it would have fallen out when Bob turned it upside down.
6. Bob walks to the garden.
- Bob left the cup in the garden, but there is no mention of the ball being removed from the cup.
7. Bob puts the cup down in the garden.
- The ball, if it was in the cup, would still be in the cup in the garden.
8. Bob walks to the garage.
- No mention of the ball being taken to the garage.
Based on the given information, we can conclude that the ball is still inside the cup in the garden.
يمكنك أن ترى كيف أدركت Chatgpt أن الكرة سقطت في الخطوة 5 ولكن لا تزال لم تحصل على الإجابة الصحيحة.
هل يمكننا القيام بأي شيء أفضل ، دون اللجوء إلى ChatGPT 4؟
شهدت الأسابيع الأخيرة ظهور العديد من الأوراق ومستودعات بيثون التي تقدم مفهوم شجرة الشجرة (TOT):
المبدأ الأساسي هو توفير فرصة LLM لاستكشاف فروع متعددة من التفكير مع التقييم الذاتي مع تقدمها. وفقًا للأوراق المشار إليها ، يبدو أن هذا النهج فعال للغاية بالنسبة لمشاكل معينة.
يبدو أن الأوراق الحالية التي تناقش إطار شجرة الفكر تعتمد على إجراء مكالمات متعددة إلى LLM ، عادةً عبر Python. واحدة من المزايا الرئيسية لسلسلة التفكير في التفكير هو أنه يمكن تحقيقه مع chatgpt في موجه جملة واحدة.
هذا دفعني إلى التفكير: هل يمكننا تحسين نتائج COT من خلال استخدام مطالبة على غرار TOTMY؟
ها هي المطالبة التي استخدمتها:
Imagine three different experts are answering this question.
All experts will write down 1 step of their thinking,
then share it with the group.
Then all experts will go on to the next step, etc.
If any expert realises they're wrong at any point then they leave.
The question is...
وهنا استجابة من chatgtp 3.5:
Expert 1: The ball is in the living room.
Expert 2: The ball is in the kitchen.
Expert 3: The ball is in the bedroom.
Expert 1: Bob carries the cup to the bedroom, so the ball must be in the cup.
Expert 2: Oh, I see my mistake. Yes, the ball is in the cup.
Expert 3: Agreed, the ball is in the cup in the bedroom.
Expert 1: Next, Bob turns the cup upside down in the bedroom.
Expert 2: After that, Bob walks to the garden and puts the cup down.
Expert 3: Therefore, the ball must have fallen out of the cup when Bob turned it upside down in the bedroom. So, the ball is in the bedroom, not in the cup anymore.
Expert 1: Oh, you're right. I made a mistake. The ball is in the bedroom, not in the cup.
Expert 2: Agreed, the ball is in the bedroom.
Expert 3: Bob then walks to the garage, so the ball remains in the bedroom. It is not in the garden or the garage.
Expert 1: Absolutely, the ball is still in the bedroom.
Expert 2: Yes, the ball hasn't moved from the bedroom.
Expert 3: Therefore, the ball is in the bedroom, not in the garden or the garage.
All three experts agree that the ball is in the bedroom.
يمكنك أن ترى كيف يحصل chatgpt 3.5 على هذا الوقت!
هناك مجال لتحسين المطالبة لتحسين الأداء ، مما قد يؤدي إلى استجابات أكثر إيجازًا.
إذا وجدت نجاحًا مع أي مطالبات أخرى ، فإننا نرحب بكم لتقديم العلاقات العامة وإضافتها إلى tree-of-thought-prompts.txt .
يشبه إطار الأشجار المفكر تشابهًا لعملية صنع القرار التنظيمية التي تشمل العديد من أصحاب المصلحة. يحاول الهيكل محاكاة الإجماع وتنوع الفكر.
يمكن تمديد التأثير باستخدام عوامل متعددة يتم ضبطها بشكل دقيق للتعامل مع أجزاء مختلفة من المهمة. يعد Autogpt مثالًا جيدًا على ذلك ، حيث يخطط LLM ذاتيًا مقابل هدف. باستخدام LLMs المضبوطة بشكل مختلف للتخصص أو حتى LLMs مختلفة تمامًا سيسمح بتنوع أكبر من التفكير. يمكن أن تتفاعل هذه LLMS في أوضاع مختلفة (باستخدام مصطلحات طبولوجيا الفريق):
غالبًا ما تتفوق الفرق ذات الأداء العالي على الأفراد في صنع القرار. لذلك ، من المعقول أن تبني الهياكل والخصائص التنظيمية الأخرى يمكن أن يعزز أداء LLMs. بالإضافة إلى تنوع الفكر والتخصص و contensus ، قد نكون قادرين على محاكاة:
يرجى الاستشهاد بهذا المستودع إذا كنت تستخدم الرمز.
@misc{tree-of-thought-prompting,
title = {Using Tree-of-Thought Prompting to boost ChatGPT's reasoning},
author = {Dave Hulbert},
year = 2023,
month = may,
journal = {GitHub repository},
publisher = {Zenodo},
doi = {10.5281/ZENODO.10323452},
url = {https://doi.org/10.5281/zenodo.10323452},
howpublished = {url{https://github.com/dave1010/tree-of-thought-prompting}}
}