AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

Towards Better Correctness and Efficiency in Code Generation

Yunlong Feng, Yang Xu, Xiao Xu, Binyuan Hui, Junyang Lin

AAAI 2026 · Reinforcement learning › RL from feedback · approach: Scaling RL with verifiable rewards

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
fine-tune with SFT on the fastest correct code only (ablation)Qwen2.5-Coder-Instruct-7B · MercuryBeyond: 74.80 -> 66.36Hurt
fine-tune with SFT on the fastest correct code only (ablation)Qwen2.5-Coder-Instruct-7B · MercuryPass@1: 77.20 -> 83.20Improved
use DPO instead of SFTQwen2.5-Coder-Instruct-7B · EvalPerfDPSnorm: 77.85 -> 87.61Improved
raise proportion of efficiency pairs in DPO data from 0.0 to 0.5 (ablation)DPO on Qwen2.5-Coder-Instruct-7B · MercuryPass@1: 87.11 -> 85.47Hurt
use only efficiency pairs (proportion 1.0) in DPO data (ablation)DPO on Qwen2.5-Coder-Instruct-7B · MercuryPass@1: 85.47 -> 77.30Hurt
use high-complexity synthetic inputs for efficiency rewards in online RL (ablation)GRPO on Qwen2.5-Coder-Instruct-7B · EvalPerfDPSnorm: 81.94 -> 83.96Improved
use high-complexity synthetic inputs for efficiency rewards in online RL (ablation)GRPO on Qwen2.5-Coder-Instruct-7B · MercuryPass@1: 90.70 -> 87.50Hurt
use RLOO instead of GRPO for online RLQwen2.5-Coder-Instruct-7B · MercuryPass@1: 87.50 -> 89.92Improved
use RLOO instead of GRPO for online RLQwen2.5-Coder-Instruct-7B · MercuryBeyond: 79.15 -> 76.00Hurt
initialize online RL from the base model instead of the instruct model (ablation)online RL · EvalPerfDPSnorm: 83.96 -> 90.81Improved
initialize online RL from DPO model with 0.1 efficiency pairs instead of the instruct model (ablation)online RL · MercuryPass@1: 87.50 -> 88.83Improved
add RLOO efficiency stage after DPO stage (two-stage tuning) (ablation)Qwen2.5-Coder-Instruct-7B · MercuryBeyond: 76.66 -> 79.04Improved

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