Contextual override, the tendency of large language models to shift away from otherwise correct predictions under misleading prompt evidence, poses a fundamental risk to reliable deployment in context-rich environments. Despite its prevalence, the internal mechanism behind this shift remains poorly understood, particularly the causal structure underlying the override process. To bridge this gap, we formulate contextual override as an internal mediation process that attributes evidence-induced answer-preference shifts to fine-grained model components. Building on this formulation, we introduce a theoretically grounded mediated-effect framework that disentangles evidence sensitivity from causal mediation, allowing override-mediating components to be precisely identified and their causal role verified through targeted ablation. We empirically validate our framework across six models and three benchmarks, demonstrating consistent support for our mediation account across diverse architectures and task settings. Our study reveals that *contextual override is routed through sparse, structurally localized pathways, rather than being diffusely entangled with the model’s internal knowledge*. These findings provide a mechanistic diagnostic for auditing contextual override and a targeted interventional test for validating the identified pathways, rather than a general-purpose defense mechanism.

Visual evidence selection is a critical component of multimodal retrieval-augmented generation (RAG), yet existing methods typically rely on semantic relevance or surface-level similarity, which are often misaligned with the actual utility of visual evidence for downstream reasoning. We reformulate multimodal evidence selection from an information-theoretic perspective by defining evidence utility as the information gain induced on a model’s output distribution. To overcome the intractability of answer-space optimization, we introduce a latent notion of evidence helpfulness and theoretically show that, under mild assumptions, ranking evidence by information gain on this latent variable is equivalent to answer-space utility. We further propose a training-free, surrogate-accelerated framework that efficiently estimates evidence utility using lightweight multimodal models. Experiments on MRAG-Bench and Visual-RAG across multiple model families demonstrate that our method consistently outperforms state-of-the-art RAG baselines while achieving substantial reductions in computational cost. We release our code at https://github.com/Hcnaeg/utility-mrag.

Real-world vision-language applications demand varying levels of perceptual granularity. However, most existing visual large language models (VLLMs), such as LLaVA, pre-assume a fixed resolution for downstream tasks, which leads to subpar performance. To address this problem, we first conduct a comprehensive and pioneering investigation into the resolution preferences of different vision-language tasks, revealing a correlation between resolution preferences with (1) image complexity, and (2) uncertainty variance of the VLLM at different image input resolutions. Building on this insight, we propose an empirical formula to determine the optimal resolution for a given vision-language task, accounting for these two factors as the zeroth-order and first-order terms in the Taylor expansion on a given image input. Second, based on rigorous experiments, we propose a novel parameter-efficient fine-tuning technique to extend the visual input resolution of pre-trained VLLMs to the identified optimal resolution. Extensive experiments on various vision-language tasks validate the effectiveness of our method.

Ad hoc dataset retrieval has become an important way of finding data on the Web, where the underlying problem is how to measure the relevance of a dataset to a query. State-of-the-art solutions for this task are still lexical methods, which cannot capture semantic similarity. Semantics-aware knowledge-enhanced retrieval methods, which achieved promising results on other tasks, have yet to be systematically studied on this specialized task. To fill the gap, in this paper, we present an empirical investigation of the task where we implement and evaluate, on two test collections, a set of implicit and explicit knowledge-enhancement retrieval methods in various settings. Our results reveal the unique features of the task and suggest an interpolation of different kinds of methods as the current best practice.