ReAdapt improves warm‑introduction and reaction selection accuracy for Gemini‑3‑Flash
The paper “When LLM Agents Fail to Read the Room: ReAdapt for Relational Social Reasoning” argues that standard LLM‑agent loops ignore relational evidence, causing surface‑obvious choices when content and relationship cues diverge. To expose this failure, the authors build a benchmark of 500 synthetic social worlds that generate 1,000 queries across two tasks—reaction selection and warm…
Key points
- Benchmark includes 500 synthetic social worlds and 1,000 queries across reaction selection and warm introduction tasks
- ReAdapt adds an explicit relational state and policy‑driven Adapt step to the ReAct loop
- On 150‑query subsets, Gemini‑3‑Flash with ReAdapt improves warm‑introduction accuracy from 37% to 51% and reaction‑selection from 69% to 77%
They introduce ReAdapt (Relationship‑Adaptive Agent with Policy‑driven sTate), which augments the ReAct loop with an explicit structured social state z = (G, B, R, N, D) capturing goal, belief, relationship, norm, and disclosure. After each tool observation, ReAdapt runs a typed Adapt step that updates this state and emits a policy operation (continue, switch, abandon, or clarify) before selecting the next action.
Evaluated with Gemini‑3‑Flash on a stratified subset of n = 150 queries per task, ReAdapt raises warm‑introduction accuracy from 37% to 51% (+14 points) and reaction‑selection accuracy from 69% to 77% (+8 points). Oracle regret drops from 0.260 to 0.152 and from 0.095 to 0.053, suggesting explicit relational‑state adaptation helps LLM agents revise decisions based on social evidence.
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Research
All →- Semantic Routing Calibration mitigates LLM over-refusal · 1 src
- AIBuildAI-2.5 ranks first on MLE-Bench with 73.3% medal rate · 1 src
- Researchers fine-tune 406M model for meeting summaries with retrieved text spans · 1 src
- Researchers test how language models handle numerical formats in word problems · 1 src
- Author pretrains language model End-to-End in Rust for $164 · 1 src
Comments
via GitHub Discussions