Open machine learning engineering tasks require agents to interpret data, build models, debug code, and produce valid competition submissions. Although current coding agents can generate runnable programs, they often rediscover solutions across tasks and repeat engineering mistakes across feedback rounds. We propose EEMA, an experience-evolving meta-agent framework that organizes accumulated experience by purpose and routes it during multi-round solving. Its static component distills public competition material and historical failure traces into compact Skills: task-level modeling priors for improving solution quality and failure-prevention contracts for maintaining execution validity. Its dynamic component, Experience Routed Search (ERS), operates above a lower-level coding agent. At each round, ERS selects a draft, debug, or improve action, routes the corresponding Skill, and updates task memory, allowing cross-task experience to guide code generation while task-level feedback shapes later decisions. On MLE-bench Lite (22 tasks), EEMA achieves a 77.3% medal rate, 20.1% average leaderboard Grade, and 96.0% execution success, outperforming all baselines including GPT-5.5-xhigh. Ablations show that task Skills improve solution quality, failure-prevention Skills improve reliability, and ERS coordinates both effectively. The materials will be released at https://github.com/FrontisAI/OpenRSI.

Enterprise agents increasingly operate in workspaces, read heterogeneous files, use tools, and deliver business artifacts. We introduce EnterpriseClawBench, an internal enterprise agent benchmark constructed from proprietary agent sessions. Starting from a large archive of real workplace sessions, the EnterpriseClawBench pipeline produces 852 reproducible tasks with recovered fixtures, rewritten prompts, role classes, skill subclasses, hard rules, and semantic rubrics. Because the data contain internal enterprise content, the benchmark data are not released; the reusable contribution is the construction and evaluation protocol. On a manually audited 120-task Lite subset, the best score is only 0.663 for Codex with GPT-5.5. The results show that enterprise agent evaluation must report harness–model combinations, artifact delivery, visual quality, cost, runtime, and skill-transfer behavior rather than a single score.

Large Language Models (LLMs) excel at general language tasks but struggle in specialized domains. Specialized Generalist Models (SGMs) address this by preserving broad capabilities while adapting to target domains. However, existing architectures provide limited support for task-guided specialized memory mechanisms. In this work, we introduce Nirvana, an SGM featuring specialized memory, linear-time complexity, and test-time task information extraction. Central to Nirvana are: (1) Task-Aware Memory Trigger (Trigger), which treats each input as a self-supervised fine-tuning task and adjusts task-related parameters on the fly; and (2) Specialized Memory Updater (Updater), which dynamically consolidates task-relevant context. Nirvana matches or surpasses LLM baselines on general benchmarks and achieves the lowest perplexity across specialized domains including biomedicine, finance, and law. On the challenging task of Magnetic Resonance Imaging (MRI), we attach lightweight codecs to the frozen Nirvana backbone and fine-tune them on paired k-space signals and images. Nirvana achieves higher-fidelity reconstructions than conventional LLM-based models, with Trigger providing effective domain-specific adaptation. Ablation studies confirm that removing Trigger leads to substantial degradation across all tasks, underscoring its essential role in task-aware specialization. Models are available at https://huggingface.co/collections/YuhuaJiang/nirvana. Code is available at https://github.com/YuhuaJiang2002/Nirvana.

Large language models are successful in answering factoid questions but are also prone to hallucination.We investigate the phenomenon of LLMs possessing correct answer knowledge yet still hallucinating from the perspective of inference dynamics, an area not previously covered in studies on hallucinations.We are able to conduct this analysis via two key ideas.First, we identify the factual questions that query the same triplet knowledge but result in different answers. The difference between the model behaviors on the correct and incorrect outputs hence suggests the patterns when hallucinations happen.Second, to measure the pattern, we utilize mappings from the residual streams to vocabulary space.We reveal the different dynamics of the output token probabilities along the depths of layers between the correct and hallucinated cases. In hallucinated cases, the output token’s information rarely demonstrates abrupt increases and consistent superiority in the later stages of the model.Leveraging the dynamic curve as a feature, we build a classifier capable of accurately detecting hallucinatory predictions with an 88% success rate. Our study shed light on understanding the reasons for LLMs’ hallucinations on their known facts, and more importantly, on accurately predicting when they are hallucinating.

Length extrapolation algorithms based on Rotary position embedding (RoPE) have shown promising results in extending the context length of language models. However, understanding how position embedding can capture longer-range contextual information remains elusive. Based on the intuition that different dimensions correspond to different frequency of changes in RoPE encoding, we conducted a dimension-level analysis to investigate the correlation between a hidden dimension of an attention head and its contribution to capturing long-distance dependencies. Using our correlation metric, we identified a particular type of attention heads, which we named Positional Heads, from various length-extrapolated models. These heads exhibit a strong focus on long-range information interaction and play a pivotal role in long input processing, as evidence by our ablation. We further demonstrate the correlation between the efficiency of length extrapolation and the extension of the high-dimensional attention allocation of these heads. The identification of Positional Heads provides insights for future research in long-text comprehension.

Insight gradually becomes a crucial form of long-term memory for an agent. However, the emergence of irrelevant insight and the lack of general insight can greatly undermine the effectiveness of insight. To solve this problem, in this paper, we introduce Multi-Scale Insight Agent (MSI-Agent), an embodied agent designed to improve LLMs’ planning and decision-making ability by summarizing and utilizing insight effectively across different scales. MSI achieves this through the experience selector, insight generator, and insight selector. Leveraging a three-part pipeline, MSI can generate task-specific and high-level insight, store it in a database, and then use relevant insight from it to aid in decision-making. Our experiments show that MSI outperforms another insight strategy when planning by GPT3.5. Moreover, We delve into the strategies for selecting seed experience and insight, aiming to provide LLM with more useful and relevant insight for better decision-making. Our observations also indicate that MSI exhibits better robustness when facing domain-shifting scenarios.