RLVF: Learning from Verbal Feedback without Overgeneralization

• URL: http://arxiv.org/abs/2402.10893v1
• Date: Fri, 16 Feb 2024 18:50:24 GMT
• Title: RLVF: Learning from Verbal Feedback without Overgeneralization
• Authors: Moritz Stephan, Alexander Khazatsky, Eric Mitchell, Annie S Chen, Sheryl Hsu, Archit Sharma, Chelsea Finn
• Abstract summary: 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.
• Score: 94.19501420241188
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The diversity of contexts in which large language models (LLMs) are deployed requires the ability to modify or customize default model behaviors to incorporate nuanced requirements and preferences. A convenient interface to specify such model adjustments is high-level verbal feedback, such as "Don't use emojis when drafting emails to my boss." However, while writing high-level feedback is far simpler than collecting annotations for reinforcement learning from human feedback (RLHF), we find that simply prompting a model with such feedback leads to overgeneralization of the feedback to contexts where it is not relevant. We study the problem of incorporating verbal feedback without such overgeneralization, inspiring a new method Contextualized Critiques with Constrained Preference Optimization (C3PO). C3PO uses a piece of high-level feedback to generate a small synthetic preference dataset specifying how the feedback should (and should not) be applied. It then fine-tunes the model in accordance with the synthetic preference data while minimizing the divergence from the original model for prompts where the feedback does not apply. Our experimental results indicate that our approach effectively applies verbal feedback to relevant scenarios while preserving existing behaviors for other contexts. For both human- and GPT-4-generated high-level feedback, C3PO effectively adheres to the given feedback comparably to in-context baselines while reducing overgeneralization by 30%.

Related papers

• Aligning Language Models Using Follow-up Likelihood as Reward Signal [40.388526412214276]
  We propose using the likelihood of follow-up utterances as rewards to differentiate preferred responses from less favored ones. Our proposed reward mechanism, Follow-up Likelihood as Reward" (FLR), matches the performance of strong reward models trained on large-scale human or GPT-4 annotated data.
  arXiv  Detail & Related papers  (2024-09-20T23:47:25Z)
• Alternate Preference Optimization for Unlearning Factual Knowledge in Large Language Models [2.0962367975513496]
  Machine unlearning aims to efficiently eliminate the influence of specific training data, known as the forget set, from the model. Existing unlearning methods rely solely on negative feedback to suppress responses related to the forget set. We propose a novel approach called Alternate Preference Optimization (AltPO), which combines negative feedback with in-domain positive feedback on the forget set.
  arXiv  Detail & Related papers  (2024-09-20T13:05:07Z)
• Beyond Thumbs Up/Down: Untangling Challenges of Fine-Grained Feedback for Text-to-Image Generation [67.88747330066049]
  Fine-grained feedback captures nuanced distinctions in image quality and prompt-alignment. We show that demonstrating its superiority to coarse-grained feedback is not automatic. We identify key challenges in eliciting and utilizing fine-grained feedback.
  arXiv  Detail & Related papers  (2024-06-24T17:19:34Z)
• Aligning Large Language Models from Self-Reference AI Feedback with one General Principle [61.105703857868775]
  We propose a self-reference-based AI feedback framework that enables a 13B Llama2-Chat to provide high-quality feedback. Specifically, we allow the AI to first respond to the user's instructions, then generate criticism of other answers based on its own response as a reference. Finally, we determine which answer better fits human preferences according to the criticism.
  arXiv  Detail & Related papers  (2024-06-17T03:51:46Z)
• Constructive Large Language Models Alignment with Diverse Feedback [76.9578950893839]
  We introduce Constructive and Diverse Feedback (CDF) as a novel method to enhance large language models alignment. We exploit critique feedback for easy problems, refinement feedback for medium problems, and preference feedback for hard problems. By training our model with this diversified feedback, we achieve enhanced alignment performance while using less training data.
  arXiv  Detail & Related papers  (2023-10-10T09:20:14Z)
• System-Level Natural Language Feedback [83.24259100437965]
  We show how to use feedback to formalize system-level design decisions in a human-in-the-loop-process. We conduct two case studies of this approach for improving search query and dialog response generation. We show the combination of system-level and instance-level feedback brings further gains.
  arXiv  Detail & Related papers  (2023-06-23T16:21:40Z)
• Learning to Simulate Natural Language Feedback for Interactive Semantic Parsing [30.609805601567178]
  We propose a new task simulating NL feedback for interactive semantic parsing. We accompany the task with a novel feedback evaluator. Our feedback simulator can help achieve comparable error correction performance as trained using the costly, full set of human annotations.
  arXiv  Detail & Related papers  (2023-05-14T16:20:09Z)
• Training Language Models with Language Feedback at Scale [50.70091340506957]
  We introduce learning from Language Feedback (ILF), a new approach that utilizes more informative language feedback. ILF consists of three steps that are applied iteratively: first, conditioning the language model on the input, an initial LM output, and feedback to generate refinements. We show theoretically that ILF can be viewed as Bayesian Inference, similar to Reinforcement Learning from human feedback.
  arXiv  Detail & Related papers  (2023-03-28T17:04:15Z)

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.