Proposes Monotonic Inference Policy Improvement (MIPI) and a two-step Monotonic Inference Policy Update (MIPU) to address training–inference probability mismatch in LLM reinforcement learning by constructing sampler-referenced candidate updates and accepting synchronized updates using an inference-gap proxy; shows improved reasoning accuracy and stability under FP8-quantized rollouts.
AI Dataset2026
Provides 300 adversarial "metacognitive-trap" problems to measure whether LLMs notice and recover from their own reasoning errors. Combines multiple-choice vulnerability tests with free-form adapter-gain evaluation and ships per-model metacognition adapters for frozen-base probing.
AI Agent Papers2026
Matches detection paradigms to four stratified attack-surface layers of AI agents — infrastructure, protocol/tool, agent behavior, and model — and presents AI-Infra-Guard: an open-source red-teaming framework with rule-based infra scanning, LLM-driven audits of MCP servers and skill packages, and a jailbreak/attack-operator harness.
AI Model2026
Provides quantized GGUF weights and configs for Agents‑A1 — a 35B Mixture-of-Experts agent trained for long-horizon, tool-enabled reasoning; supports 262K-context serving and runtimes like vLLM and SGLang.
AI Agent Papers2026
Evaluates how long-term memory in LLM agents amplifies sycophantic behavior and when memory should or should not influence decisions. Provides five targeted tasks, 1,550 standardized samples, an evaluation pipeline, and baseline adapters to test memory use, conflicts, scope, updates, and personalization.
Zhilin Wang, Han Song +14University of Science and Technology of China, The Chinese University of Hong Kong +6
Provides a benchmark and protocol to evaluate agents that iteratively edit executable policies under a fixed interaction budget, recording full execution–feedback–revise trajectories. Built from compact RL environments with trajectory-level diagnostics and hidden held-out validation.
Embodied AI2026
Provides a systematic benchmark and design roadmap for video-based world models to evaluate robot policies, introducing WMBench and GigaWorld-1 optimized for long-horizon, action-faithful rollouts. Offers controlled comparisons across model families, action encodings, and 324k+ simulated vs real rollouts, with code, models, and datasets released for reproducible evaluation.
AI Agent Papers2026
Introduces a bounded-memory, typed-retrieval contract for long-horizon LLM agents and evaluates it in Slay the Spire 2 — assembling per-decision prompts from five typed slots rather than appending raw transcripts. Key outputs include ablationable memory layers, 298 labeled trajectories, and reproducible analysis scripts.
AI Agent Papers2026
Provides a comprehensive benchmark to evaluate LLM-based data agents on realistic, multi-domain data-science workflows. Features skill-level ground-truth labels, 15 vertical domains (including real B2B tasks), and LLM-driven task generation to ensure coverage; includes an open testbed and agent evaluations.
AI Agent Papers2026
Proposes SkillOpt-Lite, a minimal pipeline for optimizing LLM agent skills by treating rollout traces as filesystem files and applying trajectory exploration, consensus mining, and independent validation; integrates as a one-line VSCode Copilot command and reports cross-benchmark improvements that let smaller models sometimes outperform larger ones.
AI Dataset2026
Converts long or messy model reasoning traces into concise, user-facing summaries with optional metadata. 61K cleaned English samples in JSON format, Apache-2.0 licensed, created to train and evaluate reasoning-summarization models and to present safe, readable explanations instead of raw chain-of-thought.
Embodied AI2026
Evaluates generalist robot manipulation policies across simulation and real-world settings using 42 sim tasks and 18 real tasks; measures generalization, memory, precision, long-horizon execution and open-vocabulary instruction following, and provides a cloud-accessible real-world evaluation system with XPolicyLab integration and a public leaderboard.