AI Paper Atlas

Research map / Multi-agent systems and games / Human-AI and goal-conditioned

Shoot First, Ask Questions Later? Building Rational Agents that Explore and Act Like People

Gabriel Grand, Valerio Pepe, Joshua B. Tenenbaum, Jacob Andreas

ICLR 2026 · Multi-agent systems and games › Human-AI and goal-conditioned · approach: Theory-of-mind social reasoning

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
answer via Python code generation instead of direct answering (ablation)15 LMs as Spotter · SpotterQAaccuracy: +13.2% absoluteImproved
combine chain-of-thought with code generation (ablation)15 LMs as Spotter · SpotterQAaccuracy: +14.7% absoluteImproved
add Bayes-Q (EIG-maximizing question selection) (ablation)Llama-4-Scout, GPT-4o Captains · CaptainQAexpected information gain: up to +0.227 bitsImproved
add Bayes-Q question selection (ablation)Llama-4-Scout, GPT-4o Captains · CaptainQAtargeting F1: +0.021 to +0.026 F1No clear effect
add Bayes-M move selection (hit-probability MAP) (ablation)weaker LM Captains · CaptainQAtargeting F1Improved
add Bayes-D one-step lookahead decision (ablation)Llama-4-Scout · CaptainQAquestion timingImproved
add full Bayes-QMD strategy (ablation)Llama-4-Scout · CaptainQAtargeting F1: 0.367 -> 0.764Improved
add full Bayes-QMD strategy (ablation)GPT-4o · CaptainQAtargeting F1: 0.450 -> 0.782Improved
add Bayesian question and move selection (ablation)GPT-5 · CaptainQAtargeting F1No clear effect
add QM Bayes strategies on Guess Who?GPT-4o · Guess Who?success rate: 0.617 -> 0.900Improved
add QM Bayes strategies on Guess Who?Llama-4-Scout · Guess Who?success rate: 0.300 -> 0.724Improved
increase context-dependence of questions (complex vs simple)GPT-4o Spotter · SpotterQAaccuracy: 72.8% -> 60.4%Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy