5arXiv cs.CL (Computation and Language)·
CurateEvo: Failure-Driven Dynamic Data Curation for Agentic LLM Post-Training
CurateEvo is a new framework for agentic post-training that treats data curation as a dynamic, evolving process rather than a fixed preprocessing step. The system represents curation strategies as executable code and iteratively rewrites them based on failed trajectories from a held-out development set, producing SFT data, RL data, and an inference-time memory bank. Evaluated on ACEBench-Agent, BFCL-V4, and τ²-Bench, CurateEvo outperforms prior curation methods by 3.2 and 2.7 average points in labeled and wild-data settings respectively, while also reducing curation overhead.
Related guides (3)
Related events (8)
Provenance-grounded gating and adaptive recovery improve synthetic post-training data curation
A controlled study examines two underexplored practices in synthetic post-training data pipelines: grounding filtering signals in source provenance and systematically recovering rejected samples rather than discarding them. Using adversarially injected corpora as ground-truth failure labels, the authors find that exact source provenance improves faithfulness gating for stronger judges, that hallucination and reward gates reject largely disjoint populations (making both necessary), and that adaptive recovery via failure diagnosis and targeted regeneration outperforms naive resampling. Generator scale is the primary driver of downstream fine-tuning quality, with filtration and recovery contributing meaningfully but secondarily.
OpenThoughts-Agent: Open data curation pipeline for broadly capable agentic models
The OpenThoughts-Agent (OT-Agent) project releases a fully open data curation pipeline for training agentic language models, addressing the gap left by prior efforts (SWE-Smith, SERA, Nemotron-Terminal) that target single benchmarks. The team conducts over 100 controlled ablation experiments and assembles a 100K-example training set, fine-tuning Qwen3-32B to achieve 44.8% average accuracy across seven agentic benchmarks — a 3.9 percentage point improvement over the strongest existing open agentic model (Nemotron-Terminal-32B at 40.9%). Training data, pipeline, experimental data, and models are publicly released at openthoughts.ai.
EEVEE: Multi-dataset test-time prompt learning framework for self-improving LLM agents
EEVEE is a new framework enabling LLM agents to perform test-time prompt learning across heterogeneous multi-dataset task streams, addressing a gap where prior methods only handled single-dataset settings. The system uses a router to partition inputs into task clusters and assigns them to suitable prompt configurations, optimized via a router-prompt co-evolution strategy. Experiments show improvements of 10.38 and 24.32 average points over Qwen3-4B-Instruct and DeepSeek-V3.2 respectively, outperforming prior SOTA methods GEPA and ACE by up to 48.2%.
Autodata: Meta-optimized agentic data scientist for high-quality synthetic data generation
Researchers introduce Autodata, a framework that trains AI agents to act as data scientists capable of generating high-quality synthetic training and evaluation data. The method includes a meta-optimization loop (Agentic Self-Instruct) that improves the data scientist agent itself, yielding further performance gains. Experiments on CS research, legal reasoning, and mathematical reasoning tasks show improvements over classical synthetic data methods. The authors frame this as a path to converting inference compute into higher-quality training data.
AgentCL: A Rigorous Evaluation Framework for Continual Learning in Language Agents
AgentCL is a new benchmark and evaluation framework designed to rigorously assess continual learning in language agents, addressing gaps in existing benchmarks that focus on retrieval over long-context documents or use naive task streams with limited cross-task analysis. The framework constructs compositional task streams where earlier sub-solutions, evidence, or workflows are intentionally reusable in later tasks, contrasting them with naive streams to measure transfer gains. The authors also introduce MemProbe, a probing method that stores interactions, insights, and skills while filtering unreliable experiences during consolidation. Empirical results across coding, deep research, and language understanding tasks show that controlled streams better distinguish memory design quality, and that naive streams can mask memory-induced degradation.
VeriEvol: Verified data construction pipeline for scaling multimodal mathematical reasoning
VeriEvol is a new framework for scaling reinforcement learning on visual mathematical reasoning by decoupling prompt difficulty expansion from answer reliability verification. It uses a type-aware evolution module to generate harder image-grounded prompts and an HTV-Agent verifier that rejects answers only after failing to find counter-evidence. Scaling SFT data from 10K to 250K samples raises mean accuracy from 35.42 to 54.73 across five visual-math benchmarks, with an additional +3.88 cumulative gain over an un-evolved RL baseline when combined with GRPO-style training. The authors release prompts, data, models, code, and full verifier traces.
EnvFactory: Scaling Tool-Use Agents via Executable Environments Synthesis and Robust RL
EnvFactory is a fully automated framework for training tool-use LLM agents via Agentic Reinforcement Learning, addressing two key bottlenecks: scalable execution environments and realistic multi-turn training data. It autonomously constructs stateful, executable tool environments from authentic resources and synthesizes natural trajectories with implicit human intents via topology-aware sampling. Using only 85 verified environments across 7 domains, it generates 2,575 SFT and RL trajectories and improves Qwen3-series models by up to +15% on BFCLv3, +8.6% on MCP-Atlas, and +6% on conversational benchmarks, outperforming prior approaches that use 5x more environments.
EvoArena benchmark and EvoMem memory paradigm for LLM agents in dynamic environments
Researchers introduce EvoArena, a benchmark suite that evaluates LLM agents in dynamic environments by modeling changes as progressive update sequences across terminal, software, and social domains. Alongside it, they propose EvoMem, a patch-based memory paradigm that records memory evolution as structured update histories to help agents reason about environmental change. Current agents score only 39.6% average accuracy on EvoArena, while EvoMem yields consistent gains on EvoArena and also improves performance on GAIA and LoCoMo benchmarks. The work highlights a significant gap between static-benchmark performance and real-world dynamic deployment requirements.