Thought-Retriever: Don’t Just Retrieve Raw Data, Retrieve Thoughts for Memory-Augmented Agentic Systems

Tao Feng [email protected], Pengrui Han [email protected], Guanyu Lin [email protected], Ge Liu [email protected], Jiaxuan You [email protected] · 2026

Read Thought-Retriever: Don’t Just Retrieve Raw Data, Retrieve Thoughts for Memory-Augmented Agentic Systems by Tao Feng [email protected], Pengrui Han [email protected], Guanyu Lin [email protected], Ge Liu [email protected], Jiaxuan You [email protected] (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-12231.txt.

Description

Feng and four colleagues propose Thought-Retriever, a method letting language-model agents draw on arbitrarily long external data without being bound by context length or retrieved chunk counts. Instead of raw chunks, it filters past intermediate responses, stores them as thought memory, and retrieves relevant thoughts for new queries, so memory improves with use. On AcademicEval and two public datasets it beats strong baselines by at least 7.6 percent in F1 and 16 percent in win rate.

Catalog record

CREATORTao Feng [email protected], Pengrui Han [email protected], Guanyu Lin [email protected], Ge Liu [email protected], Jiaxuan You [email protected]
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE97574
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.12231

Browse the full library · About the Tower of Babel