4arXiv cs.CL (Computation and Language)·
Kontrast framework detects cross-modal knowledge inconsistencies across text, tables, and knowledge graphs
Researchers introduce Kontrast, an automatic framework for detecting and categorizing knowledge inconsistencies between Wikipedia text, tables, and Wikidata knowledge graphs. The system uses Text-to-SPARQL and LLM reasoning to compare table-based answers against KG evidence, revealing conflicts, temporal mismatches, and structural gaps. Experiments on Table-QA datasets show cross-modal inconsistencies are common and informative, with implications for RAG pipelines and LLM pre-training data quality. Code and data are publicly released alongside a benchmark for future cross-modal consistency work.
Related guides (2)
Related events (8)
Contrastive-Difference CKA reveals concept-specific structural alignment across LLM architectures
Researchers introduce CKA_Delta (contrastive-difference CKA), a training-free diagnostic that isolates concept-specific representational convergence from generic similarity across LLM architectures. The method reveals a geometric-functional universality dissociation: moderate geometric alignment coexists with near-perfect functional transfer across six concept domains and multiple architectural families. CKA_Delta also functions as an architectural outlier detector, flagging Gemma as a notable outlier (d=1.08, AUC=0.79). The work provides a practical tool for cross-architecture concept monitoring without requiring model training.
KATE framework improves LLM tool calling via experiential knowledge integration and parallel reasoning
Researchers present KATE (Knowledge-Augmented Tool Execution), a framework addressing LLM failures in multi-step tool use by systematically studying knowledge acquisition, activation, and internalization. Key findings include that instance-level experiential knowledge outperforms abstract intent-level knowledge, that expanding reasoning width via parallel sampling with aggregation beats deeper chain-of-thought, and that reinforcement learning outperforms supervised fine-tuning for knowledge internalization. KATE is evaluated on BFCL-V3 and AppWorld benchmarks, showing consistent improvements over strong baselines across model scales.
Adversarial Subspace Alignment for Robust Multimodal Knowledge Editing in MLLMs
This paper addresses the generalization gap in multimodal large language model (MLLM) knowledge editing, where edits fail to propagate across semantically equivalent visual and linguistic variations. The authors introduce Latent Adversarial Robustification (LAR), which generates adversarial but semantically coherent variants in joint latent space, and Rank-Constrained Subspace Learning (RCSL), which enforces low-rank alignment of adversarial representations at the edit layer. Together these form the ASAM framework, which formalizes robustness via knowledge units grouping semantically equivalent multimodal inputs. Empirical analysis demonstrates improved generality without sacrificing reliability or locality.
WikiVQABench: A Knowledge-Grounded Visual Question Answering Benchmark from Wikipedia and Wikidata
WikiVQABench is a new human-curated VQA benchmark that requires external knowledge beyond visual perception, constructed by combining Wikipedia images, captions, and Wikidata structured knowledge with LLM-generated question candidates reviewed by human annotators. The benchmark evaluates knowledge-intensive reasoning in vision-language models, covering 15 VLMs ranging from 256M to 90B parameters. Accuracy spans 24.7% to 75.6%, indicating meaningful discrimination across model scales. The dataset and code are publicly released.
LMs encode knowledge in task-specific parameter subsets, undermining the knowledge-base analogy
A new arXiv paper investigates whether language models satisfy the consistency property of knowledge bases — that the same fact returns consistent results regardless of query form. Behavioral and mechanistic analyses reveal that LMs encode knowledge in a task-specific manner: facts acquired on one task frequently fail to transfer to others during training, and distinct parameter subsets underlie the same fact across different tasks. The authors also show that chain-of-thought reasoning derives part of its effectiveness by engaging task-specific parameters beyond those tied to the evaluation task, with implications for factual reliability and model controllability.
TABVERSE benchmark isolates table representation effects across formats in LLMs and VLMs
TABVERSE is a new controlled multimodal benchmark that evaluates LLMs and VLMs on table understanding by holding table content fixed while varying representation format (HTML, Markdown, LaTeX, rendered images). Evaluation across three tasks—Question Answering, Structural Understanding, and Structure Reconstruction—shows that representation choice substantially affects performance, with structured text generally outperforming rendered images and HTML being the most robust text format. The benchmark addresses a gap in existing evaluations where content, format, and modality vary simultaneously, making it impossible to isolate representation effects.
Introducing ConTextual: Benchmark for Joint Text-Image Reasoning in Text-Rich Scenes
Hugging Face introduces ConTextual, a new benchmark evaluating multimodal models on their ability to jointly reason over text and images in text-rich scenes. The benchmark targets a specific capability gap where models must integrate visual and textual information simultaneously rather than treating them independently. A leaderboard accompanies the benchmark to track model progress on this task.
MATCHA: Contrastive Semantic Alignment Metric for LLM Evaluation
MATCHA is a new automatic evaluation metric for LLMs that addresses a fundamental flaw in existing metrics: both token-overlap (ROUGE) and embedding-based (BERTScore) metrics routinely assign near-identical scores to semantically contradictory texts. The metric uses a dual-view approach that rewards proximity to a gold reference while penalizing adversarially generated counterfactual contradictions. Evaluated across eight benchmarks spanning QA, summarization, NLI, and semantic similarity tasks, MATCHA outperforms 23 embedding models and achieves 18.38% and 20.82% improvements over ROUGE-L and BERTScore respectively on TruthfulQA. Code and metric are publicly released.