# Beneath the Scores: Rethinking Hallucination Evaluation for Video Understanding Models

arXiv: https://arxiv.org/abs/2609.28991 · HTML: https://mlsanity.com/paper/2609.28991/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Hallucination detection and mitigation, one of 17 directions in LLM reasoning.
- Builds on 34 of its 37 references that are in the map.
- First posted on arXiv 24 Sep 2026.

## Placed in

- LLM reasoning / [Hallucination detection and mitigation](https://mlsanity.com/lr/hallucination-detection-and-mitigation/)
- LLM agents and tool use / [Multimodal and embodied agents](https://mlsanity.com/agents/multimodal-and-embodied-agents/)

## Builds on (most cited first, 30 of 34)

- [Video-LLaMA: An Instruction-tuned Audio-Visual Language Model for Video Understanding](https://mlsanity.com/paper/2306.02858/) (2023)
- [Video-MME: The First-Ever Comprehensive Evaluation Benchmark of Multi-modal LLMs in Video Analysis](https://mlsanity.com/paper/2405.21075/) (2024)
- [Video-LLaVA: Learning United Visual Representation by Alignment Before Projection](https://mlsanity.com/paper/2311.10122/) (2023)
- [MVBench: A Comprehensive Multi-modal Video Understanding Benchmark](https://mlsanity.com/paper/2311.17005/) (2023)
- [Video-ChatGPT: Towards Detailed Video Understanding via Large Vision and Language Models](https://mlsanity.com/paper/2306.05424/) (2023)
- [Are We on the Right Way for Evaluating Large Vision-Language Models?](https://mlsanity.com/paper/2403.20330/) (2024)
- [EgoSchema: A Diagnostic Benchmark for Very Long-form Video Language Understanding](https://mlsanity.com/paper/2308.09126/) (2023)
- [LLaMA-VID: An Image is Worth 2 Tokens in Large Language Models](https://mlsanity.com/paper/2311.17043/) (2023)
- [MovieChat: From Dense Token to Sparse Memory for Long Video Understanding](https://mlsanity.com/paper/2307.16449/) (2023)
- [Video-R1: Reinforcing Video Reasoning in MLLMs](https://mlsanity.com/paper/2503.21776/) (2025)
- [LVBench: An Extreme Long Video Understanding Benchmark](https://mlsanity.com/paper/2406.08035/) (2024)
- [TimeChat: A Time-sensitive Multimodal Large Language Model for Long Video Understanding](https://mlsanity.com/paper/2312.02051/) (2023)
- [VideoAgent: Long-form Video Understanding with Large Language Model as Agent](https://mlsanity.com/paper/2403.10517/) (2024)
- [VTimeLLM: Empower LLM to Grasp Video Moments](https://mlsanity.com/paper/2311.18445/) (2023)
- [VideoAgent: A Memory-augmented Multimodal Agent for Video Understanding](https://mlsanity.com/paper/2403.11481/) (2024)
- [Video Understanding with Large Language Models: A Survey](https://mlsanity.com/paper/2312.17432/) (2023)
- [When Benchmarks are Targets: Revealing the Sensitivity of Large Language Model Leaderboards](https://mlsanity.com/paper/2402.01781/) (2024)
- [Can I Trust Your Answer? Visually Grounded Video Question Answering](https://mlsanity.com/paper/2309.01327/) (2023)
- [Deep Video Discovery: Agentic Search with Tool Use for Long-form Video Understanding](https://mlsanity.com/paper/2505.18079/) (2025)
- [VideoHallucer: Evaluating Intrinsic and Extrinsic Hallucinations in Large Video-Language Models](https://mlsanity.com/paper/2406.16338/) (2024)
- [CG-Bench: Clue-grounded Question Answering Benchmark for Long Video Understanding](https://mlsanity.com/paper/2412.12075/) (2024)
- [Vript: A Video Is Worth Thousands of Words](https://mlsanity.com/paper/2406.06040/) (2024)
- [VidHalluc: Evaluating Temporal Hallucinations in Multimodal Large Language Models for Video Understanding](https://mlsanity.com/paper/2412.03735/) (2024)
- [EventHallusion: Diagnosing Event Hallucinations in Video LLMs](https://mlsanity.com/paper/2409.16597/) (2024)
- [Video-MME-v2: Towards the Next Stage in Benchmarks for Comprehensive Video Understanding](https://mlsanity.com/paper/2604.05015/) (2026)
- [When AI Benchmarks Plateau: A Systematic Study of Benchmark Saturation](https://mlsanity.com/paper/2602.16763/) (2026)
- [LongVideoAgent: Multi-Agent Reasoning with Long Videos](https://mlsanity.com/paper/2512.20618/) (2025)
- [Breaking Down Video LLM Benchmarks: Knowledge, Spatial Perception, or True Temporal Understanding?](https://mlsanity.com/paper/2505.14321/) (2025)
- [ARGUS: Hallucination and Omission Evaluation in Video-LLMs](https://mlsanity.com/paper/2506.07371/) (2025)
- [MM-InstructEval: Zero-Shot Evaluation of (Multimodal) Large Language Models on Multimodal Reasoning Tasks](https://mlsanity.com/paper/2405.07229/) (2024)

## Abstract (arXiv)

Video understanding is increasingly performed by multi-stage LLM agents that separate temporal grounding, visual observation, and reasoning. Yet these stages are typically evaluated on different benchmarks and distributions, making it difficult to determine where hallucinations originate. We first organize existing benchmarks around these stages and show that their scores provide inconsistent diagnostic signals: stronger stage-level performance does not reliably imply lower downstream hallucination, and even benchmarks targeting the same capability can disagree. We therefore introduce a causal stage-intervention protocol that overwrites individual stages while holding the downstream task fixed. Across 60,008 runs on three video-agent architectures, we find that grounding is the dominant source of downstream error, with roughly four times the causal impact of corrupting visual observations. Successful grounding depends primarily on locating the correct region rather than precise temporal overlap, explaining why standard mIoU metrics poorly predict downstream reliability. We further find that incorrect evidence is substantially more harmful than missing evidence. Finally, auditing existing benchmarks against these interventions reveals that their scores do not reliably predict causal cascade sensitivity and can fail under distribution shift. These results motivate intervention-based, stage-aware evaluation for trustworthy video agents.