AI Paper Atlas

Research map / Large language models / Alignment and preferences

ROSETTA: Constructing Code-Based Reward from Unconstrained Language Preference

Sanjana Srivastava, Kangrui Wang, Yung-Chieh Chan, Tianyuan Dai, Manling Li, Ruohan Zhang, Mengdi Xu, Jiajun Wu, Li Fei-Fei

ICLR 2026 · Large language models › Alignment and preferences · approach: Reward model design and preference models

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 Preference Grounding module (vs passing raw preference to Staging) (ablation)ROSETTA · context-dependent preferencesalignment (satisfaction)Improved
add Staging module (vs grounded preference directly to Coding) (ablation)ROSETTA · all preference typesalignmentImproved
add domain-knowledge verification follow-up questions (ablation)ROSETTA · goal-related, behavioral, preferential preferencesalignmentImproved
add follow-up verification questions (ablation)ROSETTA · corrective preferencesalignmentNo clear effect
add Preference Grounding module (ablation)ROSETTA · optimizability (success rate)success rate >50%Hurt
use open-source Qwen models instead of gpt-4o/o1-miniROSETTA · mean best rewardmean best reward: 0.623 -> 0.741Improved
paraphrase prompts with gpt-4oROSETTA · mean best rewardmean best reward: 0.623 -> 0.658Mixed
paraphrase prompts with Gemini 2.5ROSETTA · mean best rewardmean best reward: 0.623 -> 0.584Hurt
add iterative error correction to generated reward codeROSETTA · generated reward functions that run% running: 79.35% -> 95.38%Improved
increase preference iteration from 1 to 4ROSETTA · human preference sequencessuccess rate: 94.6 -> 78.1Mixed

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