Dense process supervision supplies local feedback but leaves open how that feedback should become a policy advantage. We introduce PASS (Process Advantage Signal Shaping), an advantage-construction pipeline for group-relative policy optimization (GRPO). PASS combines separate reward-channel normalization and fusion, si...
Chao Wang, Hongtao Tian, Tao Yang et al.· 0 citations
Offline benchmarks for microservice root cause analysis (RCA) rank methods by a score pooled over several applications, and a pooled lead leaves two things unstated: on which of the systems it holds, and how the score was produced. We audit both on three public benchmark families (OpenRCA, RCAEval, and PetShop), runnin...
Doctor agents are moving beyond single-turn answer generation toward evolving clinical decision systems that acquire evidence within episodes and reuse experience across episodes. Fixed-input medical QA and many interactive benchmarks provide limited support for evaluating how episode-level decisions interact with cros...
World models are increasingly used for planning by predicting future states under candidate actions. Many planning environments, however, are naturally graph-structured, with agents, tools, skills, routes, and dependencies connected through fixed or evolving relations. Understanding how prediction errors accumulate in...
Xinyuan Song, Zekun Cai· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
Distilling reasoning capabilities from strong to weak language models typically involves imitating specific solution trajectories, effectively transferring what to answer rather than how to reason. This trajectory-level imitation encourages memorization of instance-specific steps rather than acquisition of transferable...
Tianyuan Shi, Canbin Huang, Bei Li et al.· 0 citations
Litmus is a zero-label system that designs evaluation and monitoring metrics for AI pipelines by eliciting evaluation intent from source code and targeted interrogation, and results support a shift from automatic metric implementation to automatic metric specification.
Prajjwal Gupta, Prasang Gupta, Vishal Bhutani et al.· Proceedings of the 2026 Conf...· 0 citations
Context managers summarize or evict old messages so agents can operate beyond a finite window. This creates a reliability problem: removing history that still shapes the model's computation may change later actions, while retaining everything defeats compression. We study plans because agents create them early and are...
Tool-using LLM agents make early choices about what to search for, which evidence to trust, and what to do next. A wrong early choice can steer every later tool call, so an agent may produce a coherent but incorrect result. Existing evaluation does not reveal this process: final-answer scoring comes only after completi...
While RL has become a promising tool for refining world models, existing methods largely rely on conservative rollouts near the training distribution, limiting exploration, behavioral diversity, and richer dynamic discovery. In this work, we challenge this conservative paradigm. We argue that the core limitation is not...
Production LLM agents increasingly depend on real-time search, yet native search grounding bundles retrieval policy, provider choice, evidence injection, cost, latency, and generation behavior behind a single model-provider boundary. This coupling makes grounding hard to inspect, tune, reuse, or port, and can trigger S...
As Large Language Model based agents enter autonomous scientific research, their ability to resist pseudoscience becomes increasingly important. Otherwise, such systems may rapidly generate plausible yet misleading studies that contaminate academic literature and erode trust in science. We present PseudoBench, an adver...
Xinyang Liao, Lingyu Li, Huacan Liu et al.· 0 citations
Standard accuracy metrics cannot explain why LLMs handle variable tracking but fail on semantically equivalent loops. We study an internal lifecycle of code reasoning in which models first brew the answer, making it linearly recoverable many layers before it becomes self-decodable, and then diverge into one of four res...
Siyue Chen, Yifu Guo, Yuquan Lu et al.· 0 citations
