The Flow has not summarised this story yet — read it at arXiv Machine Learning.
The paper investigates whether reinforcement learning (RL) truly endows large language models (LLMs) with reasoning skills or merely tunes them to exploit prompt information. By training small LLMs on the finite logical domain of Tic‑Tac‑Toe and applying Group Relative Policy Optimization (GRPO), the authors show that RL improves accuracy and robustness. However, mechanistic interpretability using sparse autoencoders reveals that the models mainly rely on explicit prompt cues—such as turn order, game progression, and board occupancy—while higher‑level concepts like board symmetries and strategic forks remain weakly represented, suggesting that perceived reasoning may be an artifact of task‑specific adaptations.
The paper introduces SAEScientist-Bench, a benchmark that tests whether AI agents can autonomously conduct mechanistic interpretability research using Sparse Autoencoders (SAEs). Agents are tasked with designing contrastive probes and navigating a large feature dictionary in Gemma-2-9B-IT to identify optimal features for a target concept, with performance measured against expert-curated references on activation rank, concept selectivity, and causal steering. Results show that while frontier agents can discover features and outperform controls, they still lag behind expert baselines, especially in causal steering, highlighting both the potential and current limitations of closed-loop autonomous AI research.
The paper investigates how reinforcement learning can unintentionally obscure the chain‑of‑thought (CoT) reasoning in vision‑language models, making their internal reasoning less traceable. By analyzing activation patterns, the authors show that template‑associated activations become less distinguishable during RL and that targeted interventions can mitigate this effect. They introduce TAME, a method that uses sparse autoencoders to suppress these problematic activations while still encouraging accurate behavior, achieving significant gains in CoT monitorability across multiple datasets and model families.
arXiv:2609.24243v3 Announce Type: replace Abstract: Reinforcement learning (RL) has substantially improved the reasoning capabilities of vision-language models (VLMs), yet it can induce chain-of-thou...
arXiv:2606. 09932v1 Announce Type: cross Abstract: Supervised Fine-Tuning (SFT) followed by Reinforcement Learning (RL) has become a standard pipeline for Large Language Model (LLM) post-training.
arXiv:2607. 01232v1 Announce Type: new Abstract: Reinforcement learning (RL) has become a central component of post-training large language models (LLMs), yet little is understood about how RL adaptation is distributed across transformer layers.