Tag
Proposes BIRD, a two-stage self-reasoning distillation method that bootstraps concise reasoning trajectories before on-policy training, achieving stronger accuracy-efficiency trade-offs on MATH-500 and AIME benchmarks. On Qwen3-8B, accuracy improves from 86.2% to 92.0% while average response length drops from 3,099 to 1,115 tokens.
This paper presents a task-specific two-agent architecture for multimodal question answering in the QANTA 2026 shared challenge, using GPT-4.1-mini for tossup questions with confidence calibration and GPT-4.1 for bonus questions with structured reasoning. The system achieved the highest overall leaderboard score, demonstrating the effectiveness of efficient reasoning strategies.
CAT introduces a framework that leverages model self-certainty signals to autonomously adjust reasoning length based on problem difficulty, reducing overthinking and improving inference efficiency for large reasoning models.
This technical report presents Ling-2.6 and Ring-2.6, a family of trillion-parameter models designed for efficient and instant agentic intelligence, featuring architectural upgrades like hybrid linear attention and specialized training methods including KPop reinforcement learning. All checkpoints are open-sourced.
Introduces SuCo, a two-stage training framework for Large Reasoning Models that uses the concept of Minimal Sufficient CoT to reduce reasoning tokens while improving accuracy across math, code, and science benchmarks.
This paper introduces a sleep-like consolidation mechanism for Transformer-based LLMs that periodically converts recent context into persistent fast weights in SSM blocks, clearing the KV cache to improve long-horizon reasoning without increasing inference latency.
This paper proposes HyperGuide, a method that distills reasoning progress into a hyperbolic geometric signal to guide step-by-step generation in LLMs, improving multi-step reasoning efficiency without explicit tree search.
A new 30B model matches systems 20-30x its size on popular benchmarks while using up to 95% fewer reasoning tokens than comparable agentic LLMs, achieved through a learned configurator that decides when and how to reason. Model and code are openly available.
New research introduces SR²AM, a configurator that self-regulates when to use simulative reasoning, improving efficiency and performance in LLMs.
Introduces SR²AM, a framework for efficient agentic reasoning via self-regulated simulative planning, achieving competitive performance with models 20-30x larger while using 26-95% fewer reasoning tokens.
AutoTTS is an open-source tool that uses agentic discovery to automatically find optimal test-time scaling strategies for LLMs, significantly reducing token usage and cost through replay-based evaluation.
The paper introduces LaTER, a two-stage reasoning paradigm that combines latent exploration with explicit Chain-of-Thought verification to reduce token usage and improve efficiency in large language models without sacrificing accuracy.
This paper introduces STOP (Super Token for Pruning), a lightweight method that learns to prune unpromising reasoning paths early during parallel decoding by appending learnable tokens and reading KV cache states, achieving 70% token reduction while improving performance on AIME and GPQA benchmarks.
This paper introduces ReBalance, a training-free, plug-and-play method that dynamically balances overthinking and underthinking in large reasoning models, improving efficiency and accuracy across multiple benchmarks.