Entity · paper

Grounding LLM Reasoning under Incomplete Graph Evidence

paperactiveprovisionalgrounding-llm-reasoning-under-incomplete-graph-evidence-1b2fef84·1 events·first seen Jun 30, 2026

Aliases: Grounding LLM Reasoning under Incomplete Graph Evidence

More like this (12)

Knowledge-Graph Grounding Helps LLMs Only for Out-of-Training Knowledge: A Controlled Study on Clinical Question Answering Evidence-Grounded F1 Beyond the Commitment Boundary: Probing Epiphenomenal Chain-of-Thought in Large Reasoning Models Scaling LLM Reasoning from Minimal Labels: A Semi-Supervised Framework with a Lightweight Verifier Long-context Reasoning Benchmarks Operadic consistency: a label-free signal for compositional reasoning failures in LLMs GraphPO: Graph-based Policy Optimization for Reasoning Models Clinical Evidence Graph Memory Agentic Chain-of-Thought Steering for Efficient and Controllable LLM Reasoning latent reasoning Reasoning as Pattern Matching: Shared Mechanisms in Human and LLM Everyday Reasoning EDIT: Evidence-Diagnosed Intervention Training for Rule-Faithful LLM Grading

Recent events (1)

5arXiv · cs.CL·Jun 30, 2026·source ↗

Theoretical framework for grounding LLM reasoning under incomplete knowledge graph evidence

A new arXiv preprint develops a formal theoretical framework for understanding how LLMs reason when guided by incomplete knowledge graphs. The authors introduce constructs including entity anchors, typed relation residuals, path energies, and support regions, and prove that under open-world incompleteness no hard rule can simultaneously reject all false unsupported trajectories while retaining all true-but-unobserved ones. Soft grounding is characterized as a KL-regularized deformation of the LLM prior, with hard conditioning as an infinite-penalty limit. The framework yields stability bounds under evidence perturbations and has implications for GraphRAG, KGQA, graph agents, constrained decoding, and faithful generation.

Evaluation and Benchmarking Agent and Tool Ecosystem Grounding LLM Reasoning under Incomplete Graph Evidence