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

• URL: http://arxiv.org/abs/2201.06009v1
• Date: Sun, 16 Jan 2022 10:11:37 GMT
• Title: Memory-assisted prompt editing to improve GPT-3 after deployment
• Authors: Aman Madaan, Niket Tandon, Peter Clark, Yiming Yang
• Abstract summary: We show how a (simulated) user can interactively teach a deployed GPT-3, doubling its accuracy on basic lexical tasks. Our simple idea is a first step towards strengthening deployed models, potentially broadening their utility.
• Score: 55.62352349324132
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large LMs such as GPT-3, while powerful, are not immune to mistakes, but are prohibitively costly to retrain. One failure mode is misinterpreting a user's instruction (e.g., GPT-3 interpreting "What word is similar to good?" to mean a homonym, while the user intended a synonym). Our goal is to allow users to correct such errors directly through interaction -- without retraining. Our approach pairs GPT-3 with a growing memory of cases where the model misunderstood the user's intent and was provided with feedback, clarifying the instruction. Given a new query, our memory-enhanced GPT-3 uses feedback from similar, prior queries to enrich the prompt. Through simple proof-of-concept experiments, we show how a (simulated) user can interactively teach a deployed GPT-3, doubling its accuracy on basic lexical tasks (e.g., generate a synonym) where users query in different, novel (often misunderstood) ways. In such scenarios, memory helps avoid repeating similar past mistakes. Our simple idea is a first step towards strengthening deployed models, potentially broadening their utility. All the code and data is available at https://github.com/madaan/memprompt.

Related papers

• RLVF: Learning from Verbal Feedback without Overgeneralization [94.19501420241188]
  We study the problem of incorporating verbal feedback without such overgeneralization. We develop a new method Contextualized Critiques with Constrained Preference Optimization (C3PO) Our approach effectively applies verbal feedback to relevant scenarios while preserving existing behaviors for other contexts.
  arXiv  Detail & Related papers  (2024-02-16T18:50:24Z)
• Reliability Check: An Analysis of GPT-3's Response to Sensitive Topics and Prompt Wording [0.0]
  We analyze what confuses GPT-3: how the model responds to certain sensitive topics and what effects the prompt wording has on the model response. We find that GPT-3 correctly disagrees with obvious Conspiracies and Stereotypes but makes mistakes with common Misconceptions and Controversies. The model responses are inconsistent across prompts and settings, highlighting GPT-3's unreliability.
  arXiv  Detail & Related papers  (2023-06-09T19:07:31Z)
• SelfCheckGPT: Zero-Resource Black-Box Hallucination Detection for Generative Large Language Models [55.60306377044225]
  "SelfCheckGPT" is a simple sampling-based approach to fact-check the responses of black-box models. We investigate this approach by using GPT-3 to generate passages about individuals from the WikiBio dataset.
  arXiv  Detail & Related papers  (2023-03-15T19:31:21Z)
• Optimizing Prompts for Text-to-Image Generation [97.61295501273288]
  Well-designed prompts can guide text-to-image models to generate amazing images. But the performant prompts are often model-specific and misaligned with user input. We propose prompt adaptation, a framework that automatically adapts original user input to model-preferred prompts.
  arXiv  Detail & Related papers  (2022-12-19T16:50:41Z)
• Ignore Previous Prompt: Attack Techniques For Language Models [0.0]
  We propose PromptInject, a framework for mask-based adversarial prompt composition. We show how GPT-3, the most widely deployed language model in production, can be easily misaligned by simple handcrafted inputs.
  arXiv  Detail & Related papers  (2022-11-17T13:43:20Z)
• Towards Teachable Reasoning Systems [29.59387051046722]
  We develop a teachable reasoning system for question-answering (QA) Our approach is three-fold: First, generated chains of reasoning show how answers are implied by the system's own internal beliefs. Second, users can interact with the explanations to identify erroneous model beliefs and provide corrections. Third, we augment the model with a dynamic memory of such corrections.
  arXiv  Detail & Related papers  (2022-04-27T17:15:07Z)
• Few-Shot Semantic Parsing with Language Models Trained On Code [52.23355024995237]
  We find that Codex performs better at semantic parsing than equivalent GPT-3 models. We find that unlike GPT-3, Codex performs similarly when targeting meaning representations directly, perhaps as meaning representations used in semantic parsing are structured similar to code.
  arXiv  Detail & Related papers  (2021-12-16T08:34:06Z)
• Improving scripts with a memory of natural feedback [38.81097942561449]
  We create a dynamic memory architecture with a growing memory of feedbacks about errors in the output. On a script generation task, we show empirically that the model learns to apply feedback effectively. This is a first step towards strengthening deployed models, potentially broadening their utility.
  arXiv  Detail & Related papers  (2021-12-16T07:01:28Z)
• Language Models are Few-Shot Learners [61.36677350504291]
  We show that scaling up language models greatly improves task-agnostic, few-shot performance. We train GPT-3, an autoregressive language model with 175 billion parameters, and test its performance in the few-shot setting. GPT-3 achieves strong performance on many NLP datasets, including translation, question-answering, and cloze tasks.
  arXiv  Detail & Related papers  (2020-05-28T17:29:03Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.