arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2201.06009cs.CL
收藏

展开后记忆辅助提示编辑可改善GPT-3

Memory-assisted prompt editing to improve GPT-3 after deployment

Carnegie Mellon University(卡内基梅隆大学);Allen Institute for Artificial Intelligence(艾伦人工智能研究所)

Aman Madaan, Niket Tandon, Peter Clark, Yiming Yang

摘要:

像GPT-3这样的大型LM很强大,但也会犯下人类显而易见的错误。例如,GPT-3会错误地解释“什么词与good相似?“来表示同音字,而用户想要的是同义词。我们的目标是通过用户与系统的交互来有效地纠正这些错误,但不需要重新培训,这将是非常昂贵的。我们将GPT-3与不断增长的记录案例的记忆配对,其中模型误解了用户的意图,以及用户反馈以澄清。这样的记忆使我们的系统能够根据用户对过去类似情况的错误纠正反馈,为任何新的查询产生增强的提示。在四个任务(两个词汇任务,两个先进的道德推理任务),我们展示了(模拟)用户如何可以交互式地教部署的GPT-3,大大提高其准确性的查询与不同种类的误解的GPT-3。我们的方法是朝着非常大的预训练LM的低成本效用增强迈出的一步。代码、数据和指令,用于在https://www.memprompt.com/上为新任务实现MEMPROMPT。

英文摘要:

Large LMs such as GPT-3 are powerful, but can commit mistakes that are obvious to humans. For example, GPT-3 would mistakenly interpret "What word is similar to good?" to mean a homophone, while the user intended a synonym. Our goal is to effectively correct such errors via user interactions with the system but without retraining, which will be prohibitively costly. We pair GPT-3 with a growing memory of recorded cases where the model misunderstood the user's intents, along with user feedback for clarification. Such a memory allows our system to produce enhanced prompts for any new query based on the user feedback for error correction on similar cases in the past. On four tasks (two lexical tasks, two advanced ethical reasoning tasks), we show how a (simulated) user can interactively teach a deployed GPT-3, substantially increasing its accuracy over the queries with different kinds of misunderstandings by the GPT-3. Our approach is a step towards the low-cost utility enhancement for very large pre-trained LMs. Code, data, and instructions to implement MEMPROMPT for a new task at https://www.memprompt.com/.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑