# Video-OPSD: Exploiting Privileged Visual Evidence for On-Policy Self-Distillation in Video Large Language Models

arXiv: https://arxiv.org/abs/2608.27065 · HTML: https://mlsanity.com/paper/2608.27065/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Placed in Multimodal distillation, one of 12 directions in Knowledge distillation.
- Builds on 17 of its 24 references that are in the map.
- First posted on arXiv 27 Aug 2026.

## Placed in

- Knowledge distillation / [Multimodal distillation](https://mlsanity.com/distill/multimodal-distillation/)

## Builds on (most cited first, 17 of 17)

- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Video-R1: Reinforcing Video Reasoning in MLLMs](https://mlsanity.com/paper/2503.21776/) (2025)
- [TempCompass: Do Video LLMs Really Understand Videos?](https://mlsanity.com/paper/2403.00476/) (2024)
- [VideoChat-R1: Enhancing Spatio-Temporal Perception via Reinforcement Fine-Tuning](https://mlsanity.com/paper/2504.06958/) (2025)
- [Self-Distillation Enables Continual Learning](https://mlsanity.com/paper/2601.19897/) (2026)
- [Video-MMMU: Evaluating Knowledge Acquisition from Multi-Discipline Professional Videos](https://mlsanity.com/paper/2501.13826/) (2025)
- [STAR: A Benchmark for Situated Reasoning in Real-World Videos](https://mlsanity.com/paper/2405.09711/) (2024)
- [WorldSense: Evaluating Real-world Omnimodal Understanding for Multimodal LLMs](https://mlsanity.com/paper/2502.04326/) (2025)
- [Time-R1: Post-Training Large Vision Language Model for Temporal Video Grounding](https://mlsanity.com/paper/2503.13377/) (2025)
- [Video-Holmes: Can MLLM Think Like Holmes for Complex Video Reasoning?](https://mlsanity.com/paper/2505.21374/) (2025)
- [TIP: Token Importance in On-Policy Distillation](https://mlsanity.com/paper/2604.14084/) (2026)
- [Video-OPD: Efficient Post-Training of Multimodal Large Language Models for Temporal Video Grounding via On-Policy Distillation](https://mlsanity.com/paper/2602.02994/) (2026)
- [OPSDL: On-Policy Self-Distillation for Long-Context Language Models](https://mlsanity.com/paper/2604.17535/) (2026)
- [Video-KTR: Reinforcing Video Reasoning via Key Token Attribution](https://mlsanity.com/paper/2601.19686/) (2026)
- [VISD: Enhancing Video Reasoning via Structured Self-Distillation](https://mlsanity.com/paper/2605.06094/) (2026)
- [APPO: Attention-guided Perception Policy Optimization for Video Reasoning](https://mlsanity.com/paper/2602.23823/) (2026)
- [Beyond Entropy: Correctness-Aware Advantage Shaping via Contrastive Policy Optimization](https://mlsanity.com/paper/2607.14614/) (2026)

## Built on by (most cited first, 4 of 4)

- [EGSD: Event-Grounded Self-Distillation for Streaming Video Understanding](https://mlsanity.com/paper/2609.36803/) (2026)
- [Video Evidence Indexing: Learning Where to Look from Video Previews for Token-Budgeted Long-Video Question Answering](https://mlsanity.com/paper/2610.00757/) (2026)
- [Beyond Entropy: Self-Diagnostic Multi-Role Token Optimization for Video Reasoning](https://mlsanity.com/paper/2610.03400/) (2026)
- [MeSD: Multi-Evidence Self-Distillation for VideoLLM](https://mlsanity.com/paper/2610.06342/) (2026)

## Abstract (arXiv)

On-policy self-distillation (OPSD) has recently emerged as an effective post-training paradigm that improves policy optimization through dense token-level supervision from a privileged self-teacher. Despite its promise, OPSD remains largely underexplored for Video Large Language Models (Video-LLMs). Existing methods typically construct privileged teachers by augmenting their context with additional information while keeping the primary input unchanged for both teacher and student. Video reasoning, however, offers a distinct source of privileged supervision within the primary input itself: long videos contain substantial temporal redundancy, and only a small subset of frames provides the evidence necessary to answer a question. Building on this observation, we present Video-OPSD, an OPSD framework that exploits privileged visual evidence for both self-teacher construction and knowledge transfer. First, our Evidence-Grounded Self-Teacher conditions the teacher exclusively on annotated evidence frames while the student continues to reason over the complete video. This focused visual input enables the teacher to provide more informative supervision. Second, our Evidence-Guided Token Optimization adaptively weights token-level distillation according to each reasoning token's reliance on privileged visual evidence, thereby emphasizing perceptually grounded reasoning. Experiments across video understanding and reasoning benchmarks show that Video-OPSD consistently improves upon Standard OPSD across multiple backbones and achieves performance comparable to GRPO while requiring substantially less training time, establishing an effective and efficient post-training approach for Video-LLMs.