mlsanityData to 14 Oct 2026

BACAM: Behavior-Aware Continual Agent Merging for Multi-Turn Interaction

arXiv 2610.04966 ↗First posted 4 Oct 2026cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

Model merging offers a way to integrate the capabilities of specialized experts, but existing agent merging methods typically require all of them to be available at once. We study continual agent merging, which integrates incoming experts sequentially without retaining previously merged experts. Yet merging in parameter space or feature subspaces does not ensure that the merged model acquires an incoming expert's behavior on interaction trajectories. Moreover, updates toward a new expert can disrupt the merged model's previously integrated interactive behavior. Therefore, we propose Behavior-Aware Continual Agent Merging (BACAM), which learns parameter-wise merging gates from candidate-generated trajectories using expert-guided behavioral supervision. Task-level stability-plasticity control and tensor-level conflict-aware update budgets limit interference with existing capabilities while allowing new ones to be acquired. The learned gates are folded into the model weights without additional inference-time parameters. Across four interactive tasks - web shopping, tool use, information retrieval, and embodied interaction - BACAM achieves an average success rate of 62.82%, exceeding the strongest evaluated merging baseline by 21.69 percentage points. Our code is publicly available at https://github.com/shuaitongli/BACAM.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 31 of its 34 references that are in the map. First posted on arXiv 4 Oct 2026.

Builds on

31 of 34 references.

References · 31

  1. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,695
  2. 2024Qwen2.5 Technical Reportcited by 4,942
  3. 2022Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference timecited by 1,396
  4. 2023AgentBench: Evaluating LLMs as Agentscited by 1,267
  5. 2022WebShop: Towards Scalable Real-World Web Interaction with Grounded Language Agentscited by 1,251
  6. 2020ALFWorld: Aligning Text and Embodied Environments for Interactive Learningcited by 1,205
  7. 2022Editing Models with Task Arithmeticcited by 1,120
  8. 2023On-Policy Distillation of Language Models: Learning from Self-Generated Mistakescited by 939
  9. 2021MuSiQue: Multihop Questions via Single-hop Question Compositioncited by 916
  10. 2023TIES-Merging: Resolving Interference When Merging Modelscited by 818
  11. 2023Language Models are Super Mario: Absorbing Abilities from Homologous Models as a Free Lunchcited by 694
  12. 2025Group-in-Group Policy Optimization for LLM Agent Trainingcited by 497
19 more
  1. 2025ToolRL: Reward is All Tool Learning Needscited by 362
  2. 2023AgentTuning: Enabling Generalized Agent Abilities for LLMscited by 338
  3. 2024Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunitiescited by 254
  4. 2023AdaMerging: Adaptive Model Merging for Multi-Task Learningcited by 252
  5. 2026A Survey of On-Policy Distillation for Large Language Modelscited by 163
  6. 2025ReSearch: Learning to Reason with Search for LLMs via Reinforcement Learningcited by 161
  7. 2024Task Singular Vectors: Reducing Task Interference in Model Mergingcited by 147
  8. 2025Agent Lightning: Train ANY AI Agents with Reinforcement Learningcited by 109
  9. 2025No Task Left Behind: Isotropic Model Merging with Common and Task-Specific Subspacescited by 96
  10. 2025Whoever Started the Interference Should End It: Guiding Data-Free Model Merging via Task Vectorscited by 62
  11. 2025Scaling Agent Learning via Experience Synthesiscited by 44
  12. 2025Merging Models on the Fly Without Retraining: A Sequential Approach to Scalable Continual Model Mergingcited by 27
  13. 2025GEM: A Gym for Agentic LLMscited by 23
  14. 2026Behavior Knowledge Merge in Reinforced Agentic Modelscited by 11
  15. 2025Ask and Remember: A Questions-Only Replay Strategy for Continual Visual Question Answeringcited by 9
  16. 2025Null-Space Filtering for Data-Free Continual Model Merging: Preserving Stability, Promoting Plasticitycited by 8
  17. 2025Expert Merging: Model Merging with Unsupervised Expert Alignment and Importance-Guided Layer Chunkingcited by 8
  18. 2026ARM: Role-Conditioned Neuron Transplantation for Training-Free Generalist LLM Agent Mergingcited by 3
  19. 2026M2A: Synergizing Mathematical and Agentic Reasoning in Large Language Modelscited by 1

Built on by

Not cited yet.