2023
Reflexion: Language Agents with Verbal Reinforcement Learning
Cheap, and it works. When an agent fails a task, a written critique plus one retry often beats a bigger model.
Generalises chain-of-thought into a search tree. The model proposes multiple partial solutions, evaluates them against each other, and can back out of a branch that is not working, rather than committing to one linear line of reasoning.
@inproceedings{yao2023tree,
author = {Shunyu Yao and Dian Yu and Jeffrey Zhao and Izhak Shafran and Thomas L. Griffiths and Yuan Cao and Karthik Narasimhan},
title = {Tree of Thoughts: Deliberate Problem Solving with Large Language Models},
booktitle = {Advances in Neural Information Processing Systems 36 (NeurIPS 2023)},
year = {2023},
eprint = {2305.10601},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2305.10601}
}Ejentic did not author this paper. Credit belongs to the authors named above; the note is ours.
2023
Cheap, and it works. When an agent fails a task, a written critique plus one retry often beats a bigger model.
2022
The pattern behind most of the agents we build. The trace is not decoration: it is the only way to debug an agent that took a wrong turn three steps ago.
2023
The benchmark we reach for when a client asks whether an agent can be trusted with a real repository. Passing tests is a much sharper signal than "the diff looks reasonable".