AI Paper Atlas

Research map / Large language models / Agents, code and math

RECAST: Expanding the Boundaries of LLMs' Complex Instruction Following with Multi-Constraint Data

Zhengkang Guo, Wenhao Liu, Mingchen Xie, Jingwen Xu, Zisu Huang, Muzhao Tian, Jianhan Xu, Yuanzhe Shen, Qi Qian, Muling Wu, Xiaohua Wang, Heda Wang and 4 more

ICLR 2026 · Large language models › Agents, code and math · approach: Agent instruction-following benchmarks · 19 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
fine-tune on RECAST-30K instead of Tülu 3 Persona IF (SFT)Qwen2.5-7B · RECAST-Test averageOSR (%): 26.83 -> 31.25Improved
fine-tune on RECAST-30K instead of Tülu 3 Persona IF (SFT)Llama-3.1-8B · RECAST-Test averageOSR (%): 26.96 -> 30.54Improved
add RLVC reinforcement learning after SFT (ablation)Llama-3.1-8B · RECAST-Test averageOSR (%): 30.54 -> 31.96Improved
add RLVC reinforcement learning after SFT (ablation)Qwen2.5-7B · RECAST-Test averageOSR (%): 31.25 -> 32.33Improved
add RLVC after SFT (ablation)Llama-3.1-8B and Qwen2.5-7B · RECAST-Test Level 1 OSROSR (%): Llama 18.50 -> 17.00; Qwen 14.50 -> 16.50Mixed
add RLVC after SFT (ablation)Qwen2.5-7B · IFEval and FollowBenchaccuracy / HSR: FollowBench 59.60 -> 63.23Improved
add RLVC after SFT (ablation)Llama-3.1-8B · FollowBenchHSR: 57.10 -> 61.76Improved
fine-tune on RECAST-30K instead of Tülu 3 Persona IFQwen2.5-7B · IFEvalPr.(L) accuracy: 73.94 -> 73.38No clear effect
SFT on RECAST-30K instead of inference-time DVR on Instruct modelLlama-3.1-8B · RECAST-Test RSR averageRSR (%): 7.0 -> 12.5Improved
SFT on RECAST-30K instead of inference-time DVR on Instruct modelQwen2.5-7B · RECAST-Test RSR averageRSR (%): 6.5 -> 13.0Improved
fine-tune on RECAST-30K instead of Tülu 3 Persona IFQwen2.5-7B · GPQA+MuSR averageaccuracy (%): 37.60 -> 34.97Hurt
add RLVC after SFT (ablation)Qwen2.5-7B · GPQA+MuSR averageaccuracy (%): 34.97 -> 38.39Improved

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