reflexion
reflexion copied to clipboard
[NeurIPS 2023] Reflexion: Language Agents with Verbal Reinforcement Learning
Hi, Thanks for your excellent work and code! I have the following two questions in hope to get your clarification. 1. May I know the correct way to reproduce the...
Hi, I was wondering if it would be possible to add compatibility with [Gymnasium](https://github.com/Farama-Foundation/Gymnasium). It is a maintained fork of openai gym and is designed as a drop-in replacement (`import...
Comparing the original HumanEval dataset with the one in your repository reveals some inconsistencies. For instance, three instances (HumanEval_32, HumanEval_38, and HumanEval_50) are missing from your version (https://github.com/noahshinn/reflexion/blob/main/programming_runs/benchmarks/humaneval-py.jsonl). Additionally, some...
Hi there, When I was inspecting the code, I found that it was using ground truth in the Evaluator part for HotpotQA. I'm wondering is that correct? Thank you!
Hi, Thanks for the great work. Unfortunately, we are unable to reproduce your results for ReAct / Reflexion on HotpotQA. E.g. You say that ReAct+gpt-3.5-turbo has a baseline accuracy of...