paper

ProvenanceGuard: Source-Aware Factuality Verification for MCP-Based LLM Agents

paperactiveprovisionalprovenanceguard-source-aware-factuality-verification-for-mcp-based-llm-agents-b6bfb068·1 events·first seen 11h ago

Aliases: ProvenanceGuard: Source-Aware Factuality Verification for MCP-Based LLM Agents

Co-occurring entities

ProvenanceGuard Model Context Protocol

More like this (12)

ProvenanceGuard Provenance Tracking in AI Compilers through the Lens of Coalgebra Multi-Component LLM Agent Agentic Chain-of-Thought Steering for Efficient and Controllable LLM Reasoning Operadic consistency: a label-free signal for compositional reasoning failures in LLMs Scaling LLM Reasoning from Minimal Labels: A Semi-Supervised Framework with a Lightweight Verifier Expert-Aware Causal Tracing of Factual Recall in Sparse MoE Language Models A History-Aware Visually Grounded Critic for Computer Use Agents Will the Agent Recuse Itself? Measuring LLM-Agent Compliance with In-Band Access-Deny Signals Which Models Are Our Models Built On? Auditing Invisible Dependencies in Modern LLMs multimodal meta-verification Role-Agent: Bootstrapping LLM Agents via Dual-Role Evolution

Recent events (1)

6arXiv · cs.CL·11h ago·source ↗

ProvenanceGuard: Source-aware factuality verification for MCP-based LLM agents

Researchers introduce ProvenanceGuard, a verifier that checks factual claims in MCP-grounded LLM agent answers against their specific source provenance rather than pooled evidence. The system decomposes answers into atomic claims, routes each to its attributed source via MCP trace metadata, and applies NLI plus token-alignment checks to detect 'cross-source conflation' — where a claim is supported somewhere but attributed to the wrong source. Evaluated on 281 medical-domain MCP-agent traces, it achieves block F1 of 0.802 and source accuracy of 0.858 on held-out data, and detects all injected attribution swaps in 50 controlled clinical probes. The work establishes source attribution as an independent factuality axis distinct from standard grounding checks.

Evaluation and Benchmarking AI Safety Research ProvenanceGuard ProvenanceGuard: Source-Aware Factuality Verification for MCP-Based LLM Agents Model Context Protocol +1 more