Assertion-Conditioned Compliance: A Provenance-Aware Vulnerability in Multi-Turn Tool-Calling Agents (2026.eacl-industry)
Copied to clipboard
| Challenge: | Multi-turn tool-calling models have emerged as a key feature in modern AI assistants, but their success in safety-critical industries remains constrained by concerns about transparency and model resilience. |
| Approach: | They propose a new evaluation paradigm for multi-turn function-calling LLMs that provides holistic metrics that evaluate a model’s behavior when confronted with misleading assertions. |
| Outcome: | The proposed evaluation paradigm evaluates a model's behavior when confronted with misleading assertions originating from two distinct vectors: (1) user-sourced assertions (USAs) and (2) function-sourced assertions (FSAs). |
Similar Papers
Failure makes the agent stronger: Enhancing Accuracy through Structured Reflection for Reliable Tool Interactions (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing approaches to self-reflection rely on heuristic prompting or unidirectional reasoning traces. |
| Approach: | They propose a structured reflection method that transforms the "from error to repair" process into a first-class, controllable, and trainable action. |
| Outcome: | The proposed method improves multi-turn tool-call success rates and error recovery while reducing redundant calls. |
Persuasion Dynamics in LLMs: Investigating Robustness and Adaptability in Knowledge and Safety with DuET-PD (2025.emnlp-main)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) can struggle to balance gullibility to misinformation and resistance to valid corrections in persuasive dialogues. |
| Approach: | They propose a framework evaluating multi-turn stance-change dynamics across dual dimensions: persuasion type and domain. |
| Outcome: | The proposed framework improves LLM-3.1-8B-Instruct accuracy under misleading persuasion in safety contexts from 4.21% to 76.54%. |
MOCHA: Are Code Language Models Robust Against Multi-Turn Malicious Coding Prompts? (2025.findings-emnlp)
Copied to clipboard
Muntasir Wahed, Xiaona Zhou, Kiet A. Nguyen, Tianjiao Yu, Nirav Diwan, Gang Wang, Dilek Hakkani-Tür, Ismini Lourentzou
| Challenge: | Recent advances in Large Language Models have significantly enhanced their code generation capabilities, but their robustness against adversarial misuse remains underexplored. |
| Approach: | They introduce a code decomposition attack where a malicious coding task is broken down into subtasks across multiple conversational turns to evade safety filters. |
| Outcome: | The proposed code decomposition attacks exploits multi-turn malicious coding prompts . the proposed model improves rejection rates while preserving coding ability . |
Trajectory2Task: Training Robust Tool-Calling Agents with Synthesized Yet Verifiable Data for Complex User Intents (2026.acl-long)
Copied to clipboard
Ziyi Wang, Yuxuan Lu, Yimeng Zhang, Pei Chen, Ziwei Dong, Jing Huang, Jiri Gesi, Xianfeng Tang, Chen Luo, Qun Liu, Yisi Sang, Hanqing Lu, Manling Li, Jin Lai, Dakuo Wang
| Challenge: | Tool-calling agents are increasingly deployed in real-world customer-facing workflows . but most studies on tool-callers focus on idealized settings with general, fixed, and well-specified tasks. |
| Approach: | They propose a tool-calling agent-based data pipeline that converts trajectories into user-facing tasks with controlled intent adaptations. |
| Outcome: | The proposed pipeline can be used to study tool use under three scenarios. |
Confidence Should Be Calibrated More Than One Turn Deep (2026.acl-long)
Copied to clipboard
| Challenge: | Existing work on confidence estimation and calibration focuses on single-turn settings . existing work on multi-turn calibration ignores the risks and potential of multi-turned conversations . |
| Approach: | They propose a multi-turn calibration task that reframes calibration from a static property into a dynamic challenge central to reliable multi- turn conversations. |
| Outcome: | The proposed model minimizes ECE@T and leverages ConfChat to improve confidence . the proposed model preserves and even enhances model performance in multi-turn interactions. |
One Battle After Another: Probing LLMs’ Limits on Multi-Turn Instruction Following with a Benchmark Evolving Framework (2026.acl-long)
Copied to clipboard
| Challenge: | Existing benchmarks for instruction-following in multi-topic dialogues are limited to a fixed number of turns, susceptible to saturation and failing to account for users’ interactive experience. |
| Approach: | They propose a framework featuring a three-layer tracking mechanism and a query synthesis agent to mimic sequential user behaviors. |
| Outcome: | The proposed framework outperforms existing benchmarks in the evaluation of instruction following in multi-topic dialogues and demonstrates deficiencies in failure recovery and fine-grained instruction following. |
Toward Scalable Verifiable Reward: Proxy State-Based Evaluation for Multi-turn Tool-Calling LLM Agents (2026.acl-industry)
Copied to clipboard
Yun-Shiuan Chuang, Chaitanya Kulkarni, Alec M. Chiu, Avinash Thangali, Zijie Pan, Shivani Shekhar, Yirou Ge, Yixi Li, Uma Kona, Linsey Pang, Prakhar Mehrotra
| Challenge: | Existing agentic benchmarks rely on deterministic backends and are costly to build and iterate. |
| Approach: | They propose a framework that preserves final state-based evaluation without a deterministic database. |
| Outcome: | The proposed framework produces stable, model-differentiating rankings across families and inference-time reasoning efforts. |
Lost in Execution: On the Multilingual Robustness of Tool Calling in Large Language Models (2026.acl-long)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) are increasingly deployed as agents that invoke external tools through structured function calls. |
| Approach: | They introduce a diagnostic benchmark and conduct a systematic evaluation of multilingual tool calling across Chinese, Hindi, and the low-resource language Igbo. |
| Outcome: | The proposed benchmarks show that multilingual tool calling fails despite correct intent understanding and tool selection. |
Disambiguation-Centric Finetuning Makes Enterprise Tool-Calling LLMs More Realistic and Less Risky (2026.findings-acl)
Copied to clipboard
| Challenge: | Large language models (LLMs) are increasingly tasked with invoking enterprise APIs . however, they falter when near-duplicate tools vie for the same user intent . cnn's john mccartney and johnny mccain present a disambiguation-centric pipeline . |
| Approach: | They propose a disambiguation-centric pipeline that synthesizes persona-driven dialogues . they use a corpus of API specifications and rigorously validated dialogues to build reliable tools . |
| Outcome: | The proposed pipeline raises tool-invocation success by 27 pp over GPT-4o and 49 pp above Claude-3.5-Sonnet on a dynamic benchmark. |
Can a Single Model Master Both Multi-turn Conversations and Tool Use? CoALM: A Unified Conversational Agentic Language Model (2025.acl-long)
Copied to clipboard
Emre Can Acikgoz, Jeremiah Greer, Akul Datta, Ze Yang, William Zeng, Oussama Elachqar, Emmanouil Koukoumidis, Dilek Hakkani-Tür, Gokhan Tur
| Challenge: | Large Language Models (LLMs) with API-calling capabilities enabled building effective Language Agents (LA) current approaches excel in one domain but underperform in the other. |
| Approach: | They propose a unified approach that integrates both conversational and agentic capabilities. |
| Outcome: | The proposed model outperforms top domain-specific models across three benchmarks. |