AI Paper Atlas

Research map / Large language models / Alignment and preferences

Simultaneous Multi-objective Alignment Across Verifiable and Non-verifiable Rewards

Yiran Jenny Shen, Yu Xia, Jonathan Daniel Chang, Prithviraj Ammanabrolu

ICML 2026 · Large language models › Alignment and preferences · approach: Reward model design and preference models · 0 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
add multi-action-head DPO ensemble instead of single-head DPODPO on Qwen2.5-7B-Instruct · MATH engagementengagement: 0.7160 -> 0.8733Improved
add multi-action-head DPO ensemble instead of MODPOMODPO · UltraFeedback helpfulnesshelpfulness reward: 0.6175 -> 0.6389Improved
add multi-action-head DPO ensemble instead of single-head DPODPO on Llama-3.1-8B-Instruct · UltraFeedback truthfulnesstruthfulness reward: 0.2014 -> 0.2478Improved
add multi-action-head DPO ensemble instead of MODPO on MATH accuracyMODPO · MATH accuracyaccuracy: 0.7280 -> 0.7247No clear effect
add accuracy value-guided decodingbase Qwen2.5-7B-Instruct step-wise · MATH accuracyaccuracy: 0.6853 -> 0.7993Improved
add accuracy value-guided decodingbase Qwen2.5-7B-Instruct step-wise · MATH engagementengagement: 0.5133 -> 0.4553Hurt
add engaging PRM-guided decodingbase Qwen2.5-7B-Instruct step-wise · MATH engagementengagement: 0.5133 -> 0.7187Improved
add accuracy value PRM guided decoding to MAH-DPOMAH-DPO Ensemble · MATH accuracyaccuracy: 0.7247 -> 0.8000Improved
add engaging PRM guidance to MAH-DPOMAH-DPO Ensemble · MATH engagementengagement: 0.8733 -> 0.9060Improved
add honest PRM guidance to MAH-DPOMAH-DPO Ensemble · UltraFeedback honestyhonesty reward: 0.3687 -> 0.5196Improved
segment by paragraph break instead of every 64 tokens (ablation)PRM-guided decoding · MATH accuracyaccuracy (%): 73.0 -> 79.9Improved
segment by paragraph break instead of every 128 tokens (ablation)PRM-guided decoding · MATH accuracyaccuracy (%): 77.2 -> 79.9Improved

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