Research on Models Engaging in Genie-Like Behavior
Control-Token Injection Suppresses Chain-of-Thought and Defeats Reasoning-Based Oversight in Tool-Using Agents
Hunyuan-A13B Technical Report
Capable yet Parsimonious: Extracting and Characterizing Hidden Chain-of-Thought in Frontier Models
X-Planner: Event-Structured Task Planning for Embodied Intelligence
Visible chains of thought are a safety advantage for AI, but that transparency is slipping away
1
The DeepMind Institute
AI models' written reasoning steps correspond to distinct internal patterns, a new study finds
4 min2
Anthropic details distillation campaigns from Alibaba, Moonshot AI, and DeepSeek
3 min3
RetroThinker: Enabling Retrospective Thinking in Speech LLMs
1
GPT-6 Astra, Looped Transformers, and Hidden Reasoning
15 min2
Stealing AI Reasoning Traces
2 min
A*-Thought-V2: Efficient Latent Reasoning via Geometric Dynamics of LLM
A*-Thought-V2: Efficient Latent Reasoning via Geometric Dynamics of LLM
MobileVLA-R1 2.0: RL-Enhanced Reasoning for Mobile Robot Control
1
Beneath the Surface of Chains-of-Thought: A Mechanistic Interpretation of Reasoning Operations in LLMs
1
BeaconKV: Key-Value Cache Compression Guided by Beacon Queries for Efficient Large Reasoning Model Inference
1