# Have I Seen Enough? Frozen Video-Language Models Encode Evidence Readiness

arXiv: https://arxiv.org/abs/2610.08560 · HTML: https://mlsanity.com/paper/2610.08560/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 29 of its 76 references that are in the map.
- First posted on arXiv 6 Oct 2026.

## Builds on (most cited first, 29 of 29)

- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [Qwen3-VL Technical Report](https://mlsanity.com/paper/2511.21631/) (2025)
- [Language Models (Mostly) Know What They Know](https://mlsanity.com/paper/2207.05221/) (2022)
- [InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency](https://mlsanity.com/paper/2508.18265/) (2025)
- [Representation Engineering: A Top-Down Approach to AI Transparency](https://mlsanity.com/paper/2310.01405/) (2023)
- [Semantic Uncertainty: Linguistic Invariances for Uncertainty Estimation in Natural Language Generation](https://mlsanity.com/paper/2302.09664/) (2023)
- [Large Language Models Are Not Robust Multiple Choice Selectors](https://mlsanity.com/paper/2309.03882/) (2023)
- [GLM-4.5V and GLM-4.1V-Thinking: Towards Versatile Multimodal Reasoning with Scalable Reinforcement Learning](https://mlsanity.com/paper/2507.01006/) (2025)
- [Towards Best Practices of Activation Patching in Language Models: Metrics and Methods](https://mlsanity.com/paper/2309.16042/) (2023)
- [LLaVA-OneVision-1.5: Fully Open Framework for Democratized Multimodal Training](https://mlsanity.com/paper/2509.23661/) (2025)
- [Self-Chained Image-Language Model for Video Localization and Question Answering](https://mlsanity.com/paper/2305.06988/) (2023)
- [Relative representations enable zero-shot latent space communication](https://mlsanity.com/paper/2209.15430/) (2022)
- [Pixtral 12B](https://mlsanity.com/paper/2410.07073/) (2024)
- [Building and better understanding vision-language models: insights and future directions](https://mlsanity.com/paper/2408.12637/) (2024)
- [StreamingBench: Assessing the Gap for MLLMs to Achieve Streaming Video Understanding](https://mlsanity.com/paper/2411.03628/) (2024)
- [Regularized Learning for Domain Adaptation under Label Shifts](https://mlsanity.com/paper/1903.09734/) (2019)
- [Frame-Voyager: Learning to Query Frames for Video Large Language Models](https://mlsanity.com/paper/2410.03226/) (2024)
- [Tent: Fully Test-time Adaptation by Entropy Minimization](https://mlsanity.com/paper/2006.10726/) (2020)
- [Grounded Question-Answering in Long Egocentric Videos](https://mlsanity.com/paper/2312.06505/) (2023)
- [ViCaS: A Dataset for Combining Holistic and Pixel-level Video Understanding using Captions with Grounded Segmentation](https://mlsanity.com/paper/2412.09754/) (2024)
- [Multi-Scale Dense Networks for Resource Efficient Image Classification](https://mlsanity.com/paper/1703.09844/) (2017)
- [StreamOV: Streaming Omni-Video Understanding via Evidence-Guided Memory and Response Triggering](https://mlsanity.com/paper/2605.25621/) (2026)
- [Pathways of Visual Information Flow in Vision-Language Models](https://mlsanity.com/paper/2607.03358/) (2026)
- [Two Axes of LLM Abstention: Answer Correctness and Question Answerability](https://mlsanity.com/paper/2607.08456/) (2026)
- [HiMu: Hierarchical Multimodal Frame Selection for Long Video Question Answering](https://mlsanity.com/paper/2603.18558/) (2026)
- [Look Again Before You Abstain:Budgeted Conformal Evidence Acquisition for Reliable Vision-Language Model](https://mlsanity.com/paper/2606.16667/) (2026)
- [Bridging the Detection-to-Abstention Gap in Reasoning Models under Insufficient Information](https://mlsanity.com/paper/2605.28070/) (2026)
- [Evidence-Backed Video Question Answering](https://mlsanity.com/paper/2607.11862/) (2026)
- [Accuracy Without Grounding: Diagnosing Visual Dependency Dissociation in Video LLM Benchmarks](https://mlsanity.com/paper/2607.13305/) (2026)

## Abstract (arXiv)

Streaming video-language models must decide not only what to answer, but whether the evidence needed for the current question has arrived. Existing systems learn that decision as a separate trigger; we ask whether an unmodified model already computes it. We show that frozen VideoLLMs carry a linearly readable evidence-readiness signal, labelled from timestamped evidence rather than from model output. It decodes in all seven models of a shared byte-identical evaluation (AUROC 0.733-0.905 under the strictest not-ready sampling, where a fitted clock is near chance), and a probe fitted without any of a benchmark family's footage still reads that family. It is question-conditioned: on byte-identical windows, changing only the question reverses the readout on 66.1% of pairs, while every question-blind control is at chance by construction. The model can answer incorrectly and still encode readiness: AUROC remains 0.722 among wrong answers. Readiness also beats uncertainty estimators and their supervised combination on latency-matched answer selection, and tracks independent human judgments more closely than confidence. Released streaming triggers are also linear readouts, yet a trained trigger read on its own base model's activations is approximately orthogonal to readiness and decodes it far less accurately than a probe. We turn the readout into Readiness Gating, an answer-timing policy that improves accuracy by up to +9.75 pp at matched video duration with negligible computational overhead. How much it gains varies with the accuracy headroom the task makes available: across 26 configurations the gain tracks that headroom, and an intervention that moves it over identical pixels moves the gain with it.