HackerPapers

arXiv:2201.06009cs.CL

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

Aman Madaan, Niket Tandon, Peter Clark, Yiming Yang

Abstract

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/.

arXiv comments
EMNLP 2022. This version updates the title to be consistent with EMNLP camera ready
Journal reference
DOI
Subjects
cs.CLComputation and Language

On Hacker News

Linked by 3 identifiable commenters in 4 comments across 3 threads, from Mar 2022 to Jan 2023.

Comments linking this paper, per year.

Discussed in comments3 threads

  1. What Is AGI-Hard 1 comment

    • I believe our common desires to establish what is and what is not factual get in the way of innovation. Talking about something isn't going to solve the problem. Doing, might: https://arxiv.org/pdf/2201.06009.pdf

      kordlessagainView on HNParent context

  2. Building a Virtual Machine Inside ChatGPT 2 comments

    • It's already possible to get some of this effect with codex. The trick is to keep appending the interaction in the prompt (to maintain a memory of sorts). For examples, you can replicate all the prompts here: https://twitter.com/yoavgo/status/1599200756631887872 with prompt +…

      axiom92View on HNParent context

    • You are right, it's already able to do some of this with codex. It's already possible to get some of this effect with codex. The trick is to keep appending the interaction in the prompt (to maintain a memory of sorts). For examples, you can replicate all the prompts here…

      axiom92View on HNParent context

  3. GPT-3 can run code 1 comment

    • "Memory assisted prompt editing to improve GPT 3 after deployment" paper: https://arxiv.org/abs/2201.06009 video: https://www.youtube.com/watch?v=gYxJEd3EUKs

      visargaView on HNParent context

Submitted to Hacker News

Connected through shared HN threads, not inferred similarity or a recommendation.