Chameleon: Plug-and-Play Compositional Reasoning with Large Language Models
Free while signed in. Answers cite the passages they came from.
First page
The curator’s take
Key pointsA framework inferring tool sequences for compositional reasoning.
01
Tool composition: LLM plans sequences of tools (Python, search, calculator, knowledge retrievers) to solve complex problems.
02
SOTA on ScienceQA: Achieves 87% accuracy on ScienceQA and 99% on TabMWP - surpassing prior specialized models.
03
Plug-and-play design: Tools can be added/removed flexibly without retraining the LLM.
04
Agent framework precursor: Influential in the agent/tool-use research direction leading to 2024 agent frameworks.
Every Monday
Get next week’s papers.
The same picks and the same summaries, in your inbox. Free, and 176 issues deep.
Subscribe on Substack