Tool use / function calling

ExpeL

ExpeL: LLM Agents Are Experiential Learners

Superseded baseline#12 of 55 most-superseded · first seen Aug 20, 2023

Superseded — cited as a baseline and beaten by newer methods

1 papers critique it · 1 beat it on benchmarks

What papers say

Verbatim critique sentences, each from a paper that cites ExpeL as a baseline.

several approaches~zhao2024expel, xu-etal-2024-enhancing-tool depend on manually curated or pre-collected task-specific demonstrations. However, this reliance not only limits scalability but also incurs substantial offline costs, making it inefficient when addressing a wide range of diverse tasks.
Self-Guided Function Calling in Large Language Models via Stepwise Experience Recall