AI Paper Atlas

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

REAL: Reading Out Transformer Activations for Precise Localization in Language Model Steering

Li-Ming Zhan, Bo LIU, Yujie Feng, Chengqiang Xie, Jiannong Cao, Xiao-Ming Wu

ICLR 2026 · Large language models › Language, safety and interpretability · approach: Task vectors and steering · 2 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
select heads with REAL instead of ITI's probe-based selectionITI · TruthfulQA (Qwen2.5-7B-Instruct)MC1: 40.51 -> 64.21 MC2; 24.48 -> 44.43 MC1Improved
select heads with REAL instead of LoFiT's selectionLoFiT · TruthfulQA (Llama2-7B-Chat)MC1: 59.56 -> 60.90Improved
add REAL head selection to ITIITI · LLM-as-judge TruthfulQA (Llama2-7B-Chat)Truth x Info: 3.40 -> 21.39Improved
add REAL head selection to ITIITI · LLM-as-judge TruthfulQA (Qwen2.5-7B-Instruct)Truth x Info: 14.82 -> 32.11Improved
add REAL layer selection to SPARESPARE · NQSWAP (Llama3-8B)Contextual EM: 77.69 -> 80.17Improved
add REAL layer selection to SPARESPARE · MACNOISE (Llama3-8B)Parametric EM: 30.72 -> 32.57Improved
use 50% of training data instead of 100% for REAL (ablation)REAL · TruthfulQA MC1 (Qwen2.5-7B)MC1: 45.59 -> 40.27Mixed
use 50% of training data instead of 100% for ITI (ablation)ITI · TruthfulQA MC1 (Qwen2.5-7B)MC1: 30.72 -> 27.13Hurt
use adaptive importance-weighted strength instead of ITI constant strength at eps=20 (ablation)REAL · TruthfulQA MC1 (Llama2-7B-Chat)MC1: 26.09 -> 42.84Improved
steer all 1024 heads with REAL instead of ITI (ablation)REAL · TruthfulQA MC1 (Llama2-7B-Chat)MC1: 17.82 -> 31.38Improved
use adaptive strength instead of fixed strength with 100 steered heads (ablation)REAL · TruthfulQA MC1 (Llama2-7B-Chat)MC1: 29.07 -> 35.62Improved
set codebook size to 32 instead of 2 (ablation)REAL · TruthfulQA MC1 (Qwen2.5-7B)MC1: 39.78 -> 45.59Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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