Causal Tracing of Object Representations in Large Vision Language Models: Mechanistic Interpretability and Hallucination Mitigation
This work addresses the interpretability and hallucination issues in LVLMs, which are crucial for improving model faithfulness and enabling downstream applications, though it is incremental as it builds on existing causal tracing methods.
The authors tackled the problem of limited mechanistic interpretability in Large Vision-Language Models (LVLMs) by introducing the Fine-grained Cross-modal Causal Tracing (FCCT) framework, which revealed that multi-head self-attentions in middle layers and feed-forward networks are critical for visual object perception, and they proposed Intermediate Representation Injection (IRI), a training-free technique that achieved state-of-the-art performance on five benchmarks while mitigating hallucination.
Despite the remarkable advancements of Large Vision-Language Models (LVLMs), the mechanistic interpretability remains underexplored. Existing analyses are insufficiently comprehensive and lack examination covering visual and textual tokens, model components, and the full range of layers. This limitation restricts actionable insights to improve the faithfulness of model output and the development of downstream tasks, such as hallucination mitigation. To address this limitation, we introduce Fine-grained Cross-modal Causal Tracing (FCCT) framework, which systematically quantifies the causal effects on visual object perception. FCCT conducts fine-grained analysis covering the full range of visual and textual tokens, three core model components including multi-head self-attention (MHSA), feed-forward networks (FFNs), and hidden states, across all decoder layers. Our analysis is the first to demonstrate that MHSAs of the last token in middle layers play a critical role in aggregating cross-modal information, while FFNs exhibit a three-stage hierarchical progression for the storage and transfer of visual object representations. Building on these insights, we propose Intermediate Representation Injection (IRI), a training-free inference-time technique that reinforces visual object information flow by precisely intervening on cross-modal representations at specific components and layers, thereby enhancing perception and mitigating hallucination. Consistent improvements across five widely used benchmarks and LVLMs demonstrate IRI achieves state-of-the-art performance, while preserving inference speed and other foundational performance.