Attention analysis shows that LLM tool selection failures occur at the readout/decision stage, not because the model fails to attend to the correct tool definition.
hub
Apigen-mt: Agentic pipeline for multi-turn data generation via simulated agent-human interplay
20 Pith papers cite this work. Polarity classification is still indexing.
20
Pith papers citing it
hub tools
citation-role summary
background 4
citation-polarity summary
roles
background 4polarities
background 4representative citing papers
SENTINEL generates targeted tasks from model failures in a Controller-Proposer-Solver loop, raising Pass^1 from 66.4 to 74.9 on Tau2-Bench Retail and outperforming standard RL.
ISE creates 23,132 execution-grounded multi-turn OS agent trajectories via intent simulation and live execution, improving agent performance on ClawEval from 19.3 to 37.7 pass@1 with Qwen3-8B.
Terminal-World is a skill-based synthesis pipeline that generates 5,723 training environments and produces Terminal-World-32B which outperforms baselines on Terminal-Bench 2.0 using only 1.2% of the data.
RecoAtlas is a benchmark that evaluates LLM recommendation agents on behavior-grounded metrics for relevance, complementarity, and diversity in addition to semantic coherence.
COVERT generates verifiable synthetic tool-use environments for RL by validated trajectory synthesis and oracle-preserving augmentations, improving tool-use accuracy on BFCL v3 and ACEBench while remaining complementary to SFT.
τ²-bench provides a Dec-POMDP-based telecom domain with compositional task generation and a tool-constrained user simulator to measure agent performance drops in dual-control versus single-control settings.
PROVE trains LLMs on multi-step tool calls using 20 live MCP servers with 343 tools, state-grounded synthesis, and adaptive efficiency rewards, delivering gains of up to 10.2 points on BFCL Multi-Turn and similar on other benchmarks.
GAIS synthesizes diverse, high-fidelity agentic tasks from real-world MCP servers and adversarial planning, outperforming LLM-only baselines on BFCL, τ²-Bench, and ACEBench with greater data efficiency.
Tool-calling evaluations for LLM agents are highly sensitive to implementation details such as random seeds and history handling, and two new techniques accelerate RL training with wall-clock speedup and no performance degradation.
FireFly inverts task synthesis by exploring real MCP servers first via pairwise tool graphs and sub-DAG sampling, then generates 5,144 verified tasks backward from outcomes to train a 4B model that matches Claude Sonnet 4.6 on tool-calling benchmarks.
CuraView detects sentence-level faithfulness hallucinations in medical discharge summaries via GraphRAG knowledge graphs and multi-agent evidence grading, achieving 0.831 F1 on critical contradictions with a fine-tuned Qwen3-14B model and 50% relative improvement over baselines.
Agent-World autonomously synthesizes verifiable real-world tasks and uses continuous self-evolution to train 8B and 14B agents that outperform proprietary models on 23 benchmarks.
ToolWeave synthesizes realistic multi-turn tool-calling dialogues via dependent workflows and parameter provenance tracking, yielding LLMs that score higher on benchmarks such as 39.75% on BFCL-V3 multi-turn versus 23.50% on prior SOTA data.
A constrained-synthesis RL method with graduated rewards for atomic validity and orchestration consistency improves LLM turn accuracy on multi-step tool benchmarks and transfers to new API sets.
This survey categorizes agentic environments for LLMs by eight attributes and domains, introduces symbolic and neural synthesis paradigms with evaluation, and outlines four agent evolution pathways plus three environment evolution paradigms.
Proposes image-bank harness and ODE closed-loop data generation to boost multimodal deep search agents, reporting average score gains from 24.9% to 39.0% on 8 benchmarks for 8B model and 30.6% to 41.5% for 30B.
Longer action horizons bottleneck LLM agent training through instability, but training with reduced horizons stabilizes learning and enables better generalization to longer horizons.
Matrix provides a peer-to-peer multi-agent system for synthetic data generation that scales to tens of thousands of workflows and delivers 2-15x higher throughput than centralized designs without quality loss.
A comprehensive review of self-evolving AI agents that improve themselves over time, organized via a framework of inputs, agent system, environment, and optimizers, with domain-specific and safety discussions.
citing papers explorer
-
SENTINEL: Failure-Driven Reinforcement Learning for Training Tool-Using Language Model Agents
SENTINEL generates targeted tasks from model failures in a Controller-Proposer-Solver loop, raising Pass^1 from 66.4 to 74.9 on Tau2-Bench Retail and outperforming standard RL.
-
ISE: An Execution-Grounded Recipe for Multi-Turn OS-Agent Trajectories
ISE creates 23,132 execution-grounded multi-turn OS agent trajectories via intent simulation and live execution, improving agent performance on ClawEval from 19.3 to 37.7 pass@1 with Qwen3-8B.
-
Terminal-World: Scaling Terminal-Agent Environments via Agent Skills
Terminal-World is a skill-based synthesis pipeline that generates 5,723 training environments and produces Terminal-World-32B which outperforms baselines on Terminal-Bench 2.0 using only 1.2% of the data.
-
Synthesize and Reward -- Reinforcement Learning for Multi-Step Tool Use in Live Environments
PROVE trains LLMs on multi-step tool calls using 20 live MCP servers with 343 tools, state-grounded synthesis, and adaptive efficiency rewards, delivering gains of up to 10.2 points on BFCL Multi-Turn and similar on other benchmarks.
-
Scaling Agentic Capabilities via Grounded Interaction Synthesis
GAIS synthesizes diverse, high-fidelity agentic tasks from real-world MCP servers and adversarial planning, outperforming LLM-only baselines on BFCL, τ²-Bench, and ACEBench with greater data efficiency.
-
CuraView: A Multi-Agent Framework for Medical Hallucination Detection with GraphRAG-Enhanced Knowledge Verification
CuraView detects sentence-level faithfulness hallucinations in medical discharge summaries via GraphRAG knowledge graphs and multi-agent evidence grading, achieving 0.831 F1 on critical contradictions with a fine-tuned Qwen3-14B model and 50% relative improvement over baselines.
-
ToolWeave: Structured Synthesis of Complex Multi-Turn Tool-Calling Dialogues
ToolWeave synthesizes realistic multi-turn tool-calling dialogues via dependent workflows and parameter provenance tracking, yielding LLMs that score higher on benchmarks such as 39.75% on BFCL-V3 multi-turn versus 23.50% on prior SOTA data.
-
Agentic Environment Engineering for Large Language Models: A Survey of Environment Modeling, Synthesis, Evaluation, and Application
This survey categorizes agentic environments for LLMs by eight attributes and domains, introduces symbolic and neural synthesis paradigms with evaluation, and outlines four agent evolution pathways plus three environment evolution paradigms.
-
Towards On-Policy Data Evolution for Visual-Native Multimodal Deep Search Agents
Proposes image-bank harness and ODE closed-loop data generation to boost multimodal deep search agents, reporting average score gains from 24.9% to 39.0% on 8 benchmarks for 8B model and 30.6% to 41.5% for 30B.
-
Matrix: Peer-to-Peer Multi-Agent Synthetic Data Generation Framework
Matrix provides a peer-to-peer multi-agent system for synthetic data generation that scales to tens of thousands of workflows and delivers 2-15x higher throughput than centralized designs without quality loss.