arXiv:2609.37104v1 Announce Type: cross
Abstract: Open-source reasoning models provide unequal access to reasoning capability across languages. When a model can solve a problem but cannot deliver a c...
By Hongyang Li, Xiao Li, Caesar Wu, Gr\'egoire Danoy, Pascal Bouvry
The paper reports that in on‑policy distillation for large language models, reasoning performance can be improved by supervising only a tiny fraction of generated tokens—sometimes just one or two tokens per reasoning trajectory, about 0.05% of all tokens. This sparse supervision consistently matches or exceeds full‑token training across nine teacher‑student setups on mathematical reasoning, and is also validated on coding reasoning, Llama models, and PPO‑based reinforcement learning with verifiable reward. The findings suggest that effective post‑training does not require token‑intensive supervision and may align more closely with natural learning processes that focus on critical reasoning steps.
By Zhishuai Liu, Xingzi Xu, Mehmet Saygin Seyfioglu, Pan Xu, Karim Bouyarmane
Mimir is a 1.78‑B parameter language model built on the Hierarchical Reasoning Model Text (HRM‑Text) architecture, which interleaves low‑level and high‑level transformer stacks. It is trained from scratch on a curated mix of 161 instruction, reasoning, tool‑use, multilingual, and synthetic datasets totaling 70.5 B tokens per epoch (103 B for the continued‑training version Mimir v1.5), using only data that meets permissibility criteria. Across 20 benchmarks in English, Danish, and Math & Code, Mimir v1.5 outperforms all other sub‑2 B models and matches or exceeds the performance of larger 9 B‑parameter models, while memorisation audits show low risk and inference‑time ablations reveal graceful degradation when adjusting low‑level stack applications.
By Peter Schneider-Kamp, Jacob Nielsen, Gianluca Barmina, Danial Namazifard, Annemette Brok Pirchert, Kenneth Enevoldsen, Lukas Galke Poech
The paper presents practical training recipes for looped language models, showing that a compute‑efficient pipeline can reduce the training budget from 7.7 T tokens to 310 B tokens while maintaining strong reasoning performance. In controlled experiments, a 1.4 B LoopLM outperforms a parameter‑matched dense model on 12 benchmarks, achieving significant gains on GSM8K, MATH, and DROP, and approaches a 3.9 B dense model at matched inference compute. The authors also provide a minimal recipe to convert pretrained dense models into looped ones, demonstrating improvements on Qwen3‑1.7B‑Base across multiple data regimes.
By Andrei Marchenko, Viacheslav Bezrukov, Oleg Kashurin, Inessa Fedorova, Dmitry Bocharov, Yuliana Shakhvalieva, Maria Tikhonova, Valerii Ternovskii
The paper introduces masked self‑distillation, a post‑training framework that trains a language model to internalize portions of its own intermediate reasoning traces. By varying the fraction of trace internalized, the authors demonstrate that models can achieve higher inference efficiency and improved task performance on math and graph‑coloring problems. Experiments on Qwen3‑4B and Qwen3‑8B show that the method generalizes well to in‑domain out‑of‑distribution cases without catastrophic forgetting, and that supervised fine‑tuning alone can reduce trace length at the expense of generalization.
By Durgesh Kalwar, Vardhan Palod, Jaya Adithya Pavuluri, Subbarao Kambhampati
arXiv:2607. 22602v1 Announce Type: new Abstract: Inference-time scaling has emerged as a powerful paradigm for improving large language model reasoning, often delivering larger gains on difficult reasoning tasks than parameter scaling alone.
By Tingxin Yang, Zefeng Wang, Mengyue Wang, Xingcheng Zhou, Yunpu Ma
arXiv:2605.31378v2 Announce Type: replace
Abstract: Large Reasoning Models (LRMs) still struggle with fine-grained translation quality estimation (QE), even with long reasoning chains. We argue that...
By Renfei Dang, Xinye Wang, Zhejian Lai, Weilu Xu, Shimin Tao, Daimeng Wei, Min Zhang, Shujian Huang
The paper studies how the distribution of information across chain‑of‑thought steps—measured by LP variance and Caveman compression ratio—relates to answer correctness in large language models. Across six model sizes, two training paradigms, and 60 GSM8K problems, it finds that lower LP variance predicts accuracy in supervised fine‑tuned models but not in reinforcement‑learning distilled ones, that models prefer more complex distributions, and that steering density at inference time harms accuracy while pre‑generation nudging and a new Entropy‑Gated Density Routing method improve it. The work demonstrates that natural density patterns are functionally important for reasoning and offers practical techniques to enhance performance.
By Tanush Kalhan, Anupam Chattopadhyay, Sriram Ranga, Amritendu Mukherjee
arXiv:2607. 16097v1 Announce Type: cross Abstract: Reinforcement learning (RL) has become central to improving large language models (LLMs) on complex reasoning tasks, yet RL post-training is largely studied in isolation from the pretraining that precedes it.
By Jingyan Shen, Ang Li, Salman Rahman, Yifan Sun, Micah Goldblum, Matus Telgarsky, Pavel Izmailov
The paper introduces Trimmed Logit-Gap SFT (TrimSFT), a token-level reweighting strategy that adjusts supervised fine-tuning loss based on the logit gap between the correct token and its strongest competitor. TrimSFT trims supervision from tokens that are either already mastered (large logit gap) or poorly supported (small or negative logit gap), focusing learning on tokens with intermediate logit gaps. Experiments on six base models across five mathematical reasoning benchmarks show that TrimSFT consistently outperforms standard SFT, achieving the best average performance on five of six models and up to +26.9 points on MATH500.
By Yaning Jia, Chunhui Zhang, Wenxuan Xu, Xingjian Diao, Xiaoyuan Wang, Soroush Vosoughi
arXiv:2608. 05541v1 Announce Type: new Abstract: Evolution Strategy (ES) is a promising alternative to gradient-based fine-tuning for resource-constrained Large Language Model (LLM) reasoning.
By Yu Gu, Zhi Zheng, Yunpeng Ba, Xialiang Tong, Mingxuan Yuan, Zhenkun Wang
arXiv:2602. 15327v2 Announce Type: replace-cross Abstract: Machine learning model performance improvements tend to arise from competition and application.
By Hanlin Zhang, Jikai Jin, Vasilis Syrgkanis, Sham Kakade