The authors introduce Librarian Agents, a two-agent framework for long-context question answering.
LLM agents learn to organize document collections before questions arrive
A reinforcement-learned builder creates reusable filesystem workspaces that help a separate answering agent navigate dispersed evidence.
Top University
Anikait Singh · Teresa Zhang · Yoonho Lee · Roshen Sanjay Nair · Aviral Kumar · Chelsea Finn
Stanford University · Carnegie Mellon University
Research Digest··2 min read
Singh et al.
Why this paper
From Stanford University and Carnegie Mellon University · Part of Agent Memory Architecture, now 4 papers
In one line
Training a builder agent with reinforcement learning to reorganize documents into a filesystem workspace improves downstream question answering and transfers to unseen domains.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ✓Limitations stated by the authors (3 noted)
- ✓Reports numbers on named benchmarks (2 benchmarks)
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§