SourcesVESPO — Stable Off-Policy LLM Training 152 HF UpvotesarXiv·high signalVariational sequence-level soft policy optimization. Stable training at 64x staleness ratio. Top paper on HuggingFace Feb 23.SourceSource pagearXiv↳ Follow the threadPolicy dependency / Stack layerMitigating Taint-Style MCP Server Vulnerabilities Without Touching Server Code (SPELLSMITH)arXivPolicy dependency / Stack layerReason Less, Verify More: Deterministic Gates Catch a Silent Policy-Violation Failure in Tool-Using AgentsarXivPolicy dependency / Stack layerLong-Horizon-Terminal-Bench: Best Frontier Agent Clears Just 15% of Real Terminal Tasks, Mean 4.3%arXiv (Tencent Hunyuan) / HuggingFace Daily PapersPolicy dependency / Update threadEvoAgentBench: A Benchmark That Isolates Whether Agents Can Turn Past Sessions Into Reusable ProceduresarXivPolicy dependency / Stack layerSelf-Guided Test-Time Training: Let the Model Pick Its Own Evidence Spans for +15% Long-Context AccuracyarXiv / HuggingFace Daily PapersPolicy dependency / Stack layerKronQ Pushes 2-Bit Quantization of LLaMA-3-70B to 7.93 Perplexity Where GPTQ Falls ApartarXiv (COLM 2026) / HuggingFace Daily PapersPolicy dependency / Stack layerExplanation-Guided Training Steers Models Toward Faithful Partial-Dependence ExplanationsarXivPolicy dependency / Stack layer'When Is Routing Meaningful?' Fewer Than 10 Agents Recover Most Diversity in Model SocietiesarXiv