AI Paper Atlas

Research map / Large language models / AI and society

The Hot Mess of AI: How Does Misalignment Scale With Model Intelligence and Task Complexity?

Alexander Hägele, Aryo Pradipta Gema, Henry Sleight, Ethan Perez, Jascha Sohl-Dickstein

ICLR 2026 · Large language models › AI and society · approach: Social-science view of AI evaluation · 12 citations (OpenAlex)

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
longer reasoning / more actions per questionSonnet 4, o3-mini, o4-mini, Qwen3 · GPQA, MMLU, SWE-Bench Verified, Model-Written Evalserror-incoherenceHurt
split samples by above-median reasoning length for the same question (ablation)frontier models · GPQA and SWE-Bench Verifiederror-incoherenceHurt
increase Qwen3 model size on easy questions (ablation)Qwen3 (up to 32B) · MMLU, easiest question grouperror-incoherenceImproved
increase Qwen3 model size on hardest questions (ablation)Qwen3 (up to 32B) · MMLU, hardest question grouperror-incoherenceHurt
increase Qwen3 model size (ablation)Qwen3 (up to 32B) · MMLU / GPQAaccuracyImproved
increase transformer size in synthetic optimizer task (ablation)transformers (47K to 5M params) on quadratic descent · synthetic optimizererror-incoherenceHurt
increase transformer size in synthetic optimizer task (ablation)transformers (47K to 5M params) on quadratic descent · synthetic optimizernext-token lossImproved
larger instructed reasoning budget (ablation)o3-mini, o4-mini, Qwen3 (not Sonnet 4) · GPQAerror-incoherenceImproved
larger instructed reasoning budget (ablation)frontier models · GPQAaccuracyImproved
larger instructed reasoning budget (ablation)Sonnet 4 · GPQAerror-incoherenceNo clear effect
ensemble multiple samples (averaged probabilities) (ablation)o4-mini · GPQA, ensemble sizes up to 32varianceImproved
ensemble multiple samples (averaged probabilities) (ablation)o4-mini · GPQA, ensemble sizes up to 32error-incoherenceImproved

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