AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Building Social World Model with Large Language Models

Haofei Yu, Yining Zhao, Guanyu Lin, Jiaxuan You

ICML 2026 · Large language models › Language, safety and interpretability · approach: Reasoning mechanisms and failure modes

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
scale world model backbone from 0.6B to 8B (ablation)SWM · Kalshi (attr, posterior mode)MASE: 0.884 -> 0.738Improved
scale world model backbone from 0.6B to 8B (ablation)SWM · Kalshi (attr, posterior mode)Corr: 0.290 -> 0.525Improved
scale world model backbone from 0.6B to 8B (ablation)SWM · Polymarket (attr, posterior mode)MASE: 0.977 -> 0.892Improved
lengthen look-back window from 1 to 16 (ablation)SWM · Kalshi (attr, posterior mode)MASE: 0.787 -> 0.738Improved
lengthen look-back window from 1 to 16 (ablation)SWM · Kalshi (attr, posterior mode)Corr: 0.430 -> 0.525Improved
scale prior attributor from 0.6B to 8B (ablation)SWM prior attributor · Kalshimedian KL to oracle posterior: 2.10 -> 1.18Improved
scale prior attributor from 0.6B to 8B (ablation)SWM prior attributor · KalshiRecall@3: 78% -> 88%Improved
use sharper Qwen3.5-397B posterior instead of Qwen3-32B (ablation)SWM training · Kalshi (attributed subset)Corr: 0.357 -> 0.525Improved
use sharper Qwen3.5-397B posterior instead of Qwen3-32B (ablation)SWM training · all-set coverageattribution coverage: 47.8% -> 11.9%Hurt
select event candidates with posterior instead of random (ablation)SWM · KalshiMASEImproved
replace prior attributor with oracle posterior at inference (ablation)SWM · KalshiDA: 0.845 -> 0.894Improved
replace prior attributor with oracle posterior at inference (ablation)SWM · Polymarket (attr)DA: 0.685 -> 0.844Improved

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