AI Paper Atlas

Research map / Large language models / Alignment and preferences

Conflict-Aware Adaptive Alignment for LLM Hallucination Mitigation

Ruohan Zong, Yang Zhang, Dong Wang

ICML 2026 · Large language models › Alignment and preferences · approach: Active and token-level preference learning

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 conflict-aware margin (apply helpfulness regulation only to conflicting pairs) (ablation)CAMP · TruthfulQAaccuracy (%): 83.48 -> 85.92Improved
add conflict-aware margin (ablation)CAMP · TruthfulQAhallucination rate (%): 11.26 -> 9.79Improved
add adaptive margin strength (per-pair margin instead of batch-wise constant) (ablation)CAMP · TruthfulQAaccuracy (%): 84.70 -> 85.92Improved
add adaptive margin strength (ablation)CAMP · TruthfulQAtruthfulness score (1-5): 4.36 -> 4.41Improved
add CAMP margin term over DPO on truthfulness pairs (ablation)DPO on truthfulness pairs · TruthfulQAhallucination rate (%): 11.26 -> 9.79Improved
increase helpfulness regulation lambda from 0 to 0.25-0.75 (ablation)CAMP · TruthfulQAhelpfulness scoreImproved
increase lambda to 1.0 (ablation)CAMP · TruthfulQAhelpfulness scoreHurt
increase lambda from 0 to 1.0 (ablation)CAMP · TruthfulQAtruthfulness and accuracyMixed
use CAMP instead of MODPO on Qwen2.5-7B-InstructCAMP · TruthfulQAhallucination rate (%): 7.22 -> 5.88Improved
replace truthfulness score with honesty score as the preference signal (ablation)CAMP · TruthfulQAhallucination rate (%): 12.48 (MODPO) -> 9.79 (CAMP)Improved
CAMP instead of CPO baselineLlama-3.1-8B-Instruct · TruthfulQAhallucination rate (%): 15.18 -> 9.79Improved
increase training pairs from 2.5k to 25k (ablation)CAMP · TruthfulQAhelpfulness, truthfulness, accuracy, hallucinationImproved

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