1
TTHE: Test-Time Harness Evolution
测试时计算新范式：TTHE方法让模型在推理阶段自我进化，无需重新训练。
arXiv:2607.08124v1 Announce Type: cross Abstract: The behavior of an LLM agent is determined not only by the underlying model, but also by its harness…
2
What Drives Interactive Improvement from Feedback?
研究自然语言反馈是否真比单纯重复尝试更有效，揭示多轮语言智能体性能提升的真实驱动力。
arXiv:2606.30774v1 Announce Type: new Abstract: We study when natural-language feedback produces improvement beyond the gains obtainable from repeated…
3
How Inference Compute Shapes Frontier LLM Evaluation
前沿论文揭示推理计算如何成为衡量LLM性能的关键，测试时计算分配正重塑评估标准。
arXiv:2606.17930v1 Announce Type: new Abstract: AI evaluations are shifting toward harder tasks that benefit from longer trajectories involving tool u…
4
Less is More: Improving LLM Reasoning with Minimal Test-Time Intervention
揭秘LLM推理效率新思路：最小干预即可显著提升性能，少即是多！
arXiv:2510.13940v4 Announce Type: replace-cross Abstract: Recent progress in large language models (LLMs) has focused on test-time scaling to improve …
5
Diversity Matters: Revisiting Test-Time Compute in Vision-Language Models
这篇论文重新审视了视觉语言模型中测试时计算的关键因素，揭示多样性对模型推理能力的重要影响。
arXiv:2605.30713v1 Announce Type: new Abstract: Test-time compute (TTC) strategies have emerged as a lightweight approach to boost reasoning in large …
6
OpenDeepThink: Parallel Reasoning via Bradley--Terry Aggregation
提出OpenDeepThink方法，用Bradley-Terry模型实现并行推理，无需真实验证器即可筛选最佳候选，为LLM推理扩展新路径。
arXiv:2605.15177v1 Announce Type: new Abstract: Test-time compute scaling is a primary axis for improving LLM reasoning. Existing methods primarily sc…