AI Paper Atlas

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

Identifying and Evaluating Inactive Heads in Pretrained LLMs

Pedro Sandoval-Segura, Xijun Wang, Ashwinee Panda, Micah Goldblum, Ronen Basri, Tom Goldstein, David Jacobs

ICLR 2026 · Large language models › Language, safety and interpretability · approach: Task vectors and steering

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
identify inactive heads with head-output-norm score functions instead of first-token attention weight (AWFT) (ablation)14 pretrained LLMs · MMLU (5-shot)% of heads zeroed within 1% accuracy: 4.61 -> 12.18Improved
use head-output-norm score instead of AWFT on Llama-3.1-8B-Instruct (ablation)Llama-3.1-8B-Instruct · MMLU (5-shot)% of heads zeroed within 1% accuracy: 1.01 -> 10.97Improved
use output-norm scores instead of AWFT on models where AWFT is already strong (ablation)Qwen2.5-14B, Qwen2.5-1.5B, Llama-3.2-3B · MMLU (5-shot)% of heads zeroed within 1% accuracy: Qwen2.5-14B 11.04 -> 9.88No clear effect
use output-norm scores instead of AWFT on PIQA (ablation)14 pretrained LLMs · PIQA (0-shot)% of heads zeroed within 1% accuracy: 10.01 -> 21.38Improved
use output-norm scores instead of AWFT on WinoGrande (ablation)14 pretrained LLMs · WinoGrande (5-shot)% of heads zeroed within 1% accuracy: 6.90 -> 14.45Improved
define head output with output projection slice (circuit-based) instead of pre-projection output (ablation)AHON (LN) · MMLU (5-shot), 3 models% of heads zeroed within 1% accuracy: 8.60 -> 8.29No clear effect
fine-tune base models (SFT, DPO, RLHF)OLMo-2 and Llama, Qwen base models · head score distributions (Wasserstein distance)similarity of attention behaviorNo clear effect
scale Qwen2.5 from 0.5B up to 14BQwen2.5 · head score distributionssimilarity of attention behaviorMixed
normalize head scores by layer average (LN) for stability across datasetsAHON vs AWFT · OLMo-2-1124-7B-Instruct thresholds on MMLU, PIQA, WinoGrandethreshold consistency for 15% heads: AWFT tau 0.077 / 0.265 / 0.109 vs AHON (LN) 0.457 / 0.435 / 0.473Improved

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