Genming Cui, Xiangyu Li, Zixu Hao, Xin Ding, Yunxin Liu, Ting Cao · Agent Quantisation Distillation
SteerDistill: Quantizing LLM Agents via Closed-Loop Teacher Steering
SteerDistill is a quantisation-aware distillation method for LLM agents, in which a full-precision teacher steers a low-bit student to revise failed responses using execution feedback. Across five agentic benchmarks and three models under W4A16 quantisation, it beats the strongest quantised baseline by 1.67 to 2.18 points in average score and stays within one point of BF16. On τ²-Bench it outperforms QeRL by 1.84 percentage points in success rate with 67% fewer training GPU hours.
Open the source ↗Similar Research & Safety releases
- Human-in-the-loop Topology Optimisation Copilot
MIT (Dat Ha, Josephine Carstensen, Ferdous Alam, Markus Buehler, Faez Ahmed) · Design Copilot · Week 47, 2026
- Read Less, Solve More: Token-Efficient Sparse Reading of External Content for AI Agents
Zedong Liu, Jiaan Wu, Xinyang Ma, Le Xu, Kai Wang, Yuanchao Hu, Dingwen Tao, Guangming Tan (Institute of Computing Technology) · Token-efficient agent reading · Week 47, 2026
- On the Design of Hyper-Connections: Simplification, Stability, and Scalability
Xiangdong Zhang et al. (dots studio) · Simpler Hyper-Connections · Week 47, 2026
- Theoretical Limits of Language Model Alignment
Apple Machine Learning Research (Lucas Monteiro Paes, Natalie Mackraz, Barry-John Theobald, Federico Danieli) · Alignment limits theory · Week 47, 2026
- The Surprising Effectiveness of Deleting Weights in LLM Post-Training
Agentic Learning AI Lab (Jack Lu, Zhenbang Yang, Mike Lasby, Tejas Pote, Yani Ioannou, Mengye Ren) · Weight Deletion Tuning · Week 47, 2026
- Incentivizing Self-Learning in Large Language Models via Reinforcement Learning
Lingkai Kong, Yuzhe Gu, Haiteng Zhao, Zijian Wu, Youqing Fang, Zhouqi Hua, Yanan Sun, Kai Chen (Intern Large Models) · RL Self-Learning · Week 47, 2026
Published by Quanta AI Labs, an AI agency in India. Browse every week in AI News Weekly.