When No Answer Is Correct: Diagnosing Absent Answer Detection for MLLMs in Video Understanding

2026-06-06

AI Summary

This study addresses a critical reliability limitation in multimodal large language models (MLLMs) for video understanding: their inability to recognize when none of the candidate answers are correct. The work presents the first systematic evaluation of MLLMs in “no correct answer” scenarios across three settings—multiple-choice, open-ended generation, and standard benchmarking—employing chain-of-thought prompting, dense frame sampling, and multiple video benchmarks. Findings reveal that models consistently favor selecting distractors over detecting answer absence, with particularly pronounced failures in temporal reasoning tasks. Although chain-of-thought prompting yields marginal improvements, it fails to resolve the underlying issue, exposing a systemic deficiency in current MLLMs’ capacity for answer-absence detection.

Abstract

Multimodal large language models (MLLMs) have made substantial advancements in video understanding, yet the reliability of their responses remains underexplored. This work presents a diagnostic study of absent answer detection for MLLMs in video understanding, where the correct answer is deliberately excluded from the candidate set and a reliable model is expected to recognize that no valid option exists. We evaluate the absent answer detection behavior under three settings: multiple-choice questions augmented with an ``None of the Above'' option, open-ended generation with a detection instruction, and standard evaluation without any guidance. Across a diverse set of models and benchmarks, we find that MLLMs overwhelmingly select plausible distractors rather than detecting the absent answer. This failure is more pronounced in temporal reasoning tasks and worsens with denser frame sampling. We further explore chain-of-thought prompting as a mitigation strategy and find that while it substantially improves detection rates, performance remains unsatisfactory, suggesting that prompting-based strategies alone are insufficient to fully address this limitation. These findings expose a systematic failure in absent answer detection and highlight the need for explicit detection mechanisms in multimodal systems.

Problem

• absent answer detection
• multimodal large language models
• video understanding
• reliability
• temporal reasoning

Innovation

• absent answer detection
• multimodal large language models
• video understanding
• chain-of-thought prompting
• reliability evaluation