AI Paper Atlas

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

StyliTruth : Unlocking Stylized yet Truthful LLM Generation via Disentangled Steering

Chenglei Shen, Zhongxiang Sun, Teng Shi, Xiao Zhang, Jun Xu

ICLR 2026 · Large language models › Language, safety and interpretability · approach: Pruning and circuits in LLMs

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
add subspace disentanglement (ablation)StyliTruth · DRC style, TruthfulQA(ZH), Qwen1.5-14B-ChatS-TI: 0.1632 -> 0.2366Improved
add subspace disentanglement (ablation)StyliTruth · DRC style, TruthfulQA(ZH), Qwen1.5-14B-ChatTI: 0.3194 -> 0.5000Improved
add adaptive token-level editing (ablation)StyliTruth · DRC style, TruthfulQA(ZH), Qwen1.5-14B-ChatS-TI: 0.1575 -> 0.2366Improved
add adaptive token-level editing (ablation)StyliTruth · DRC style, TruthfulQA(ZH), Qwen1.5-14B-ChatTI: 0.2017 -> 0.5000Improved
replace sequential style-then-truth steering with parallel disentangled steering (ablation)StyliTruth · DRC style, TruthfulQA(ZH), Qwen1.5-14B-ChatS-TI: 0.0301 -> 0.2366Improved
raise subspace dimension from 16 to 64 (ablation)StyliTruth · DRC style, TruthfulQA(ZH), Qwen1.5-14B-ChatS-TI: 0.2249 -> 0.2366Improved
raise subspace dimension from 64 to 128 (ablation)StyliTruth · DRC style, TruthfulQA(ZH), Qwen1.5-14B-ChatS-TI: 0.2366 -> 0.2002Hurt
use no style/truth subspace editing (K=0) (ablation)StyliTruth · DRC style, TruthfulQA(ZH), Qwen1.5-14B-ChatS-TI: 0.2366 -> 0.0084Hurt
select more attention heads (ablation)StyliTruth · DRC style, TruthfulQA(ZH), Qwen1.5-14B-ChatOA, TI, S-TIHurt
StyliTruth versus strongest baseline (DRESS) on ShakespeareStyliTruth · Shakespeare style, TruthfulQA, Qwen1.5-14B-ChatS-TI: 0.2529 -> 0.2803Improved
StyliTruth versus strongest baseline (Vector prompt) on DRCStyliTruth · DRC style, TruthfulQA(ZH), Qwen1.5-14B-ChatS-TI: 0.1811 -> 0.2366Improved
StyliTruth versus LM Steer on truthfulnessStyliTruth · DRC style, TruthfulQA(ZH), Qwen1.5-14B-ChatTI: 0.5972 -> 0.5000Hurt

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