Can Dynamic Clue Bottlenecks generalize to out-of-distribution robustness on VCR adversarial splits when scale

Abstract

Abstract: Recent advances in multimodal large language models (LLMs) have shown extreme effectiveness in visual question answering (VQA). However, the design nature of these end-to-end models prevents them from being interpretable to humans, undermining trust and applicability in critical domains. While post-hoc rationales offer certain insight into understanding model behavior, these explanations are not guaranteed to be faithful to the model. In this paper, we address these shortcomings by introducing an interpretable by design model that factors model decisions into intermediate human-legible explana

Research Question

Can Dynamic Clue Bottlenecks generalize to out-of-distribution robustness on VCR adversarial splits when scaled to larger backbone LLMs (e.g., 7B vs 13B parameters), measured by accuracy drop under distribution shift?

Verification Level

Descriptive public verification status only; aggregate claim counts are public, but individual claim records are not exposed here.

Quality Tier

Descriptive public triage only; this tier does not alter current publication or DOI behavior.

Quality Dimensions

Automated triage signals derived from public fields; not human peer review or independent validation.

Correction Record

Public corrections are additive records. Current status does not claim the synthesis is error-free.

Provenance