{"total":31,"items":[{"citing_arxiv_id":"2607.01942","ref_index":41,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Atomic Task Graph: A Unified Framework for Agentic Planning and Execution","primary_cat":"cs.AI","submitted_at":"2026-07-02T09:34:20+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"ATG maintains explicit DAGs of subtasks to enable dependency tracking, parallel execution, and localized repair in LLM agents, outperforming baselines on three benchmarks with 7B-8B models.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.31229","ref_index":1,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Agentic-Ideation: Sample Efficient Agentic Trajectories Synthesis for Scientific Ideation Agents","primary_cat":"cs.AI","submitted_at":"2026-06-30T07:07:51+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Agentic-Ideation uses oracle-guided multi-agent synthesis to generate efficient training trajectories for scientific ideation agents, reporting 11.91% quality gains and over 10x sample efficiency versus workflow baselines.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.27147","ref_index":5,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Safe Autoregressive Image Generation with Iterative Self-Improving Codebooks","primary_cat":"cs.CV","submitted_at":"2026-06-25T15:18:31+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Iterative self-improving codebooks enhance safety in autoregressive multimodal models by self-identifying unsafe generations and updating the codebook to eliminate harmful visual token mappings without external feedback.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.21740","ref_index":14,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Training the Orchestrator: A Supervised Approach to End-to-End PDDL Planning with LLM Agents","primary_cat":"cs.AI","submitted_at":"2026-06-19T20:53:10+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"HALO trains an orchestrator policy on verifier-approved refinement trajectories across 11 PDDL domains, matching GPT-5-mini success rates at roughly 45x lower orchestration cost and cutting LLM calls by 40-50%.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.12674","ref_index":24,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Evoflux: Inference-Time Evolution of Executable Tool Workflows for Compact Agents","primary_cat":"cs.AI","submitted_at":"2026-06-10T21:01:06+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Evoflux applies evolutionary search at inference time to repair executable tool workflows for compact agents, outperforming SFT and SFT+DPO on held-out MCP-Bench tasks with live servers and 250 tools.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.07367","ref_index":31,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Self-evolving LLM agents with in-distribution Optimization","primary_cat":"cs.LG","submitted_at":"2026-06-05T15:09:52+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Q-Evolve unifies automatic process-reward labeling via advantage estimation and behavior-proximal policy optimization inside an in-distribution RL loop to enable self-evolving LLM agents on interactive tasks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.04120","ref_index":2,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"SaliMory: Orchestrating Cognitive Memory for Conversational Agents","primary_cat":"cs.CL","submitted_at":"2026-06-02T18:31:50+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"SALIMORY trains an LM to orchestrate cognitive memory operations via stage-wise process rewards, cutting memory failures by one-third and more than doubling good personalization rates.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.02372","ref_index":42,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"COMAP: Co-Evolving World Models and Agent Policies for LLM Agents","primary_cat":"cs.AI","submitted_at":"2026-06-01T15:21:17+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"COMAP co-evolves textual world models and agent policies for LLMs through on-policy self-distillation, yielding up to 16.75% relative gains on embodied planning, web navigation, and tool-use tasks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.24893","ref_index":6,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"AgentOdyssey: Open-Ended Long-Horizon Text Game Generation for Test-Time Continual Learning Agents","primary_cat":"cs.CL","submitted_at":"2026-05-29T22:40:51+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Introduces AgentOdyssey, a procedural generator of open-ended long-horizon text games, to evaluate test-time continual learning agents and diagnose limits in exploration, memory, and planning.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.28774","ref_index":75,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Agent Explorative Policy Optimization for Multimodal Agentic Reasoning","primary_cat":"cs.CL","submitted_at":"2026-05-27T17:36:39+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"AXPO addresses the Thinking-Acting Gap in agentic RL training by targeted resampling of tool calls in all-wrong subgroups, delivering +1.8pp gains over GRPO on nine multimodal benchmarks with an 8B model beating a 32B baseline on Pass@4.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.24828","ref_index":2,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Test-Time Deep Thinking to Explore Implicit Rules","primary_cat":"cs.AI","submitted_at":"2026-05-24T02:41:44+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"TTExplore trains a 7B thinker via task-score RL to infer implicit rules at test time, raising agent success by 14-19 points on five embodied tasks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.22502","ref_index":31,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Compiling Agentic Workflows into LLM Weights: Near-Frontier Quality at Two Orders of Magnitude Less Cost","primary_cat":"cs.AI","submitted_at":"2026-05-21T13:54:11+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Compiling agentic workflows into LLM weights creates subterranean agents with near-frontier quality at two orders of magnitude less cost, validated empirically on travel booking, Zoom support, and insurance claims tasks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.17352","ref_index":1,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"AMATA: Adaptive Multi-Agent Trajectory Alignment for Knowledge-Intensive Question Answering","primary_cat":"cs.CL","submitted_at":"2026-05-17T09:45:24+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"AMATA is an adaptive multi-agent trajectory alignment system that improves factual consistency in knowledge-intensive QA via intra-trajectory preference learning and inter-agent dependency optimization.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.14133","ref_index":87,"ref_count":2,"confidence":0.9,"is_internal_anchor":false,"paper_title":"ClawForge: Generating Executable Interactive Benchmarks for Command-Line Agents","primary_cat":"cs.AI","submitted_at":"2026-05-13T21:34:08+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"ClawForge is a generator framework that creates reproducible executable benchmarks for command-line agents under state conflict, with ClawForge-Bench showing frontier models reach at most 45.3% strict accuracy and that state inspection drives most performance gaps.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.10325","ref_index":1,"ref_count":2,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Verifiable Process Rewards for Agentic Reasoning","primary_cat":"cs.AI","submitted_at":"2026-05-11T10:30:53+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"VPR converts symbolic, constraint, or posterior oracles into dense turn-level rewards for RL, improving credit assignment in agentic reasoning and transferring to general benchmarks.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"correctness can be objectively verified, supervise the reasoning process rather than only the final answer, and transfer the resulting skills to broader agentic settings-and we hope it motivates further work on verifiable environments, stronger process oracles, and methods for extending precise process supervision to less structured real-world tasks. 9 References [1] Baian Chen, Chang Shu, Ehsan Shareghi, Nigel Collier, Karthik Narasimhan, and Shunyu Yao. Fireact: Toward language agent fine-tuning.arXiv preprint arXiv:2310.05915, 2023. [2] Lang Feng, Zhenghai Xue, Tingcong Liu, and Bo An. Group-in-group policy optimization for llm agent training.arXiv preprint arXiv:2505.10978, 2025. [3] Zhibin Gou, Zhihong Shao, Yeyun Gong, yelong shen, Yujiu Yang, Nan Duan, and Weizhu"},{"citing_arxiv_id":"2605.10118","ref_index":48,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Plan in Sandbox, Navigate in Open Worlds: Learning Physics-Grounded Abstracted Experience for Embodied Navigation","primary_cat":"cs.RO","submitted_at":"2026-05-11T07:34:30+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"SAGE trains agents in physics-grounded semantic abstractions via RL with asymmetric clipping, achieving 53.21% LLM-Match Success on A-EQA (+9.7% over baseline) and encouraging physical robot transfer.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.10999","ref_index":3,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"SkillGen: Verified Inference-Time Agent Skill Synthesis","primary_cat":"cs.LG","submitted_at":"2026-05-09T19:24:11+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"SkillGen synthesizes auditable skills from agent trajectories via contrastive induction on successes and failures, then verifies net performance impact by comparing outcomes with and without the skill on identical tasks.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"learning task for automatic inference-time skill synthesis: to produce a single, auditable skill that improves a base agent. (2) We introduce SKILLGEN, a multi-agent framework that learns from both failed and successful trajectories via contrastive induction, and then generates new candidate skills that are iteratively refined and verified. The final skills are selected to have a positive net-effect on the overall performance. (3) We provide an extensive empirical study to demonstrate consistent and large held-out performance gains. SKILLGENoutperforms state-of-the-art skill-generation baselines and produces skills that transfer across models without parameter updates. 2 Preliminaries We view inference-time skills asinterventionsthat modify the behavior of a base agent and thereby"},{"citing_arxiv_id":"2605.08013","ref_index":6,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Learning CLI Agents with Structured Action Credit under Selective Observation","primary_cat":"cs.AI","submitted_at":"2026-05-08T17:02:31+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"CLI agents trained with RL benefit from selective observation via σ-Reveal and structured credit assignment via A³ that leverages AST action sub-chains and trajectory margins.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"InProceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 12248-12267, Bangkok, Thailand, 2024. [4] Anthropic. Introducing claude opus 4.7. Anthropic research announcement, 2026. [5] Max Brunsfeld and tree-sitter contributors. tree-sitter/tree-sitter: v0.25.3. Zenodo software release, 2025. Software, version 0.25.3. [6] Baian Chen, Chang Shu, Ehsan Shareghi, Nigel Collier, Karthik Narasimhan, and Shunyu Yao. Fireact: Toward language agent fine-tuning, 2023. arXiv:2310.05915 [cs.CL]. [7] DeepSeek-AI. DeepSeek-R1 incentivizes reasoning in LLMs through reinforcement learning. Nature, 645:633-638, 2025. [8] Yifeng Ding, Hung Le, Songyang Han, Kangrui Ruan, Zhenghui Jin, Varun Kumar, Zijian"},{"citing_arxiv_id":"2605.07725","ref_index":45,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"SOD: Step-wise On-policy Distillation for Small Language Model Agents","primary_cat":"cs.CL","submitted_at":"2026-05-08T13:30:42+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"SOD reweights on-policy distillation strength step-by-step using divergence to stabilize tool use in small language model agents, yielding up to 20.86% gains and 26.13% on AIME 2025 for a 0.6B model.","context_count":1,"top_context_role":"method","top_context_polarity":"use_method","context_text":"sub-billion-parameter model to reach this level on such a challenging reasoning benchmark. 2 Related Work Reinforcement Learning for Agents.RL-based post-training has evolved from reinforcement learning from human feedback (RLHF) [ 42, 43] with PPO [ 44] to more scalable methods like GRPO [20]. For language agents, structured reasoning paradigms such as ReAct [3], Toolformer [4], 2 and FireAct [45] enable tool use but rely on demonstrations rather than online optimization. Recent work extends RL to agent interaction trajectories across code generation [ 46], tool use [47], GUI interaction [48], and web navigation [ 49]. A central challenge is credit assignment under sparse, delayed feedback, addressed via trajectory-level updates and value-free formulations [50, 51]."},{"citing_arxiv_id":"2605.15207","ref_index":41,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"TeamTR: Trust-Region Fine-Tuning for Multi-Agent LLM Coordination","primary_cat":"cs.LG","submitted_at":"2026-05-01T23:42:57+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"TeamTR is a trust-region framework for multi-agent LLM fine-tuning that resamples trajectories after each update to convert quadratic compounding occupancy shift into linear scaling and yields per-update improvement lower bounds.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.27859","ref_index":6,"ref_count":3,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Rethinking Agentic Reinforcement Learning In Large Language Models","primary_cat":"cs.AI","submitted_at":"2026-04-30T13:43:25+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":3.0,"formal_verification":"none","one_line_summary":"The paper reviews conceptual foundations, methodological innovations, effective designs, critical challenges, and future directions for LLM-based Agentic Reinforcement Learning.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"fosters a powerful synergy: while reasoning facilitates the induction and updating of action plans, actions themselves interface with external environments to gather critical information. Applied to a diverse set of language and decision- making tasks, this approach [110], demonstrates clear superiority over state-of-the-art baselines, offering enhanced human interpretability compared to non-interactive methods. Extending this line of inquiry, work [6] is proposed to fine-tune LMs using trajectories from multiple tasks and prompting strategies, revealing that data diversity is key to improving agent robustness and generalization. These findings collectively establish the comprehensive benefits of agent fine-tuning, moving beyond mere scaling effects to address efficiency and cost. Addressing the limitations of"},{"citing_arxiv_id":"2604.23194","ref_index":3,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"From Coarse to Fine: Self-Adaptive Hierarchical Planning for LLM Agents","primary_cat":"cs.AI","submitted_at":"2026-04-25T07:54:23+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"AdaPlan-H enables LLM agents to generate self-adaptive hierarchical plans that adjust detail level to task difficulty, improving success rates in multi-step tasks.","context_count":1,"top_context_role":"baseline","top_context_polarity":"baseline","context_text":"against several commonly used explicit guidance- based planning methods: (1) ReAct (Yao et al., 2023b): The agent reasons before taking an action during decision-making; (2) Base Planner: A stan- dard prompting-based planning approach without training the planner, utilizing in-context learning with examples to generate plans. We use GPT-4o for its plan generation. (3) MPO (Xiong et al., 2025): After initializing the planner via SFT on meta-plans generated by LLMs, MPO samples gen- erations from the SFT model and applies the same Monte Carlo method as ours to construct prefer- ence pairs for DPO. From our perspective, MPO can be viewed as the method that only performs a single level of planning for all tasks. Actor Models."},{"citing_arxiv_id":"2604.08000","ref_index":4,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"PASK: Toward Intent-Aware Proactive Agents with Long-Term Memory","primary_cat":"cs.AI","submitted_at":"2026-04-09T09:06:13+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":4.0,"formal_verification":"none","one_line_summary":"PASK introduces the DD-MM-PAS paradigm for streaming proactive agents with intent-aware detection, hybrid memory modeling, and a new real-world benchmark where the IntentFlow model matches top LLMs on latency while finding deeper intents.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.06370","ref_index":9,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"ForkKV: Scaling Multi-LoRA Agent Serving via Copy-on-Write Disaggregated KV Cache","primary_cat":"cs.DC","submitted_at":"2026-04-07T18:52:25+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"ForkKV uses copy-on-write disaggregated KV cache with DualRadixTree and ResidualAttention kernels to deliver up to 3x throughput over prior multi-LoRA serving systems with negligible quality loss.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"generation, and testing agents, where each agent builds their own context based on previous agents. However, successfully executing such diverse subtasks requires distinct agent capabilities. A single monolithic model often lacks the flexibility to handle every stage op- timally, necessitating fine-tuning the base model with task-specific datasets to serve specialized agents effectively [9, 47, 50, 72]. To tailor foundational models for these diverse tasks in a work- flow, Parameter-Efficient Fine-Tuning (PEFT) [41] techniques, par- ticularly Low-Rank Adaptation (LoRA) [21], offer a promising solu- tion. By freezing the pretrained weights and updating only small low-rank matrices known asadapters, LoRA maintains high gener- ation quality while introducing minimal parameter overhead [15,"},{"citing_arxiv_id":"2602.20867","ref_index":43,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"SoK: Agentic Skills -- Beyond Tool Use in LLM Agents","primary_cat":"cs.CR","submitted_at":"2026-02-24T13:11:38+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"The paper systematizes agentic skills beyond tool use, providing design pattern and representation-scope taxonomies plus security analysis of malicious skill infiltration in agent marketplaces.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2511.11362","ref_index":6,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"On-Device Fine-Tuning via Backprop-Free Zeroth-Order Optimization","primary_cat":"cs.LG","submitted_at":"2025-11-14T14:46:29+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"MeZO enables larger models for on-device fine-tuning by estimating gradients via forward passes only, with theoretical size estimates and numerical results showing accuracy benefits when wall-clock time is sufficient.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2509.02547","ref_index":93,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"The Landscape of Agentic Reinforcement Learning for LLMs: A Survey","primary_cat":"cs.AI","submitted_at":"2025-09-02T17:46:26+00:00","verdict":"ACCEPT","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Survey that defines agentic RL for LLMs via POMDPs, introduces a taxonomy of planning/tool-use/memory/reasoning capabilities and domains, and compiles open environments from over 500 papers.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"demonstrate that, even when initialized from base models without any imitation traces, RL training can elicit emergent capabilities, e.g., self-correction of faulty code, adaptive adjustment of invocation frequency, and the composition of multiple tools for complex sub-tasks. Subsequently, a recent surge in research has produced works such as OTC-PO [93], ReTool [94], AutoTIR [95], VTool-R1 [96], DeepEyes [97], Pixel- Reasoner [98], Agentic Reasoning [99], ARTIST [100], ToRL [101] and numerous other works [102, 103, 104,105,106,107,108,109,110],whichemployRLpoliciesthatinterleavesymboliccomputation(e.g.,code execution, image editing) with natural-language reasoning within a single rollout. This integrated control"},{"citing_arxiv_id":"2507.02592","ref_index":2,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"WebSailor: Navigating Super-human Reasoning for Web Agent","primary_cat":"cs.CL","submitted_at":"2025-07-03T12:59:07+00:00","verdict":"CONDITIONAL","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"WebSailor trains open-source web agents to match proprietary performance on complex information-seeking tasks by generating high-uncertainty scenarios and using a new RL method called DUPO.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2506.03610","ref_index":42,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Orak: A Foundational Benchmark for Training and Evaluating LLM Agents on Diverse Video Games","primary_cat":"cs.AI","submitted_at":"2025-06-04T06:40:33+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"Orak is a foundational benchmark providing training data, interfaces, and evaluation tools for LLM agents across diverse video game genres.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2502.00955","ref_index":3,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Efficient Multi-Agent System Training with Data Influence-Oriented Tree Search","primary_cat":"cs.CL","submitted_at":"2025-02-02T23:20:16+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"DITS replaces Q-value guidance in MCTS with influence scores for synthetic data synthesis in multi-agent LLM training, claiming better efficiency and performance on eight datasets.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2401.05459","ref_index":247,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Personal LLM Agents: Insights and Survey about the Capability, Efficiency and Security","primary_cat":"cs.HC","submitted_at":"2024-01-10T09:25:45+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":3.0,"formal_verification":"none","one_line_summary":"This survey discusses key components and challenges for Personal LLM Agents and reviews solutions for their capability, efficiency, and security.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"complex task into several steps with the help of the LLM, then solve each step through either LLM inference or invoking personal tools (e.g., schedule a meeting). Sensing the context or generating the memory may also rely on the reasoning 23 Efficiency EfficientInference (§5.1) Model Compression (§5.1.1) Quantization Weight-only-Quant: GPTQ [246], AWQ [247], LLM-QAT [248], etc. Co-Quant: ZeroQuant [249], SmoothQuant [250], etc. Pruning LLM-Pruner [251], SparseGPT [252], Wanda [253], etc. Knowledge DistillationWhite-box: BabyLlama [254], MiniLLM [255], etc. Black-box: Hsieh et al. [256], SCoTD [257], etc. Low-rank FactorizationZeroQuant-V2 [258], LoSparse [259], etc. Inference Acceleration (§5.1.2) Context Compression"}],"limit":50,"offset":0}