Mixture of Probes: Learning from Privileged Modalities in Multimodal LLMs Through Probing
cs.CV, cs.LG
Submitted: 2026-07-09
Updated: 2026-11-01
Code: https://github.com/Sony/MoP
Terminology
Sources
- Qwen2.5-VL Technical Report
- X-LLM: Bootstrapping Advanced Large Language Models by Treating Multi-Modalities as Foreign Languages
- Qwen2.5 Technical Report
- From My View to Yours: Learning Egocentric Cues from Exocentric Video using Privileged Egocentric Supervision
- TimeChat: A Time-sensitive Multimodal Large Language Model for Long Video Understanding
- LLaMA: Open and Efficient Foundation Language Models
- Qwen2.5-Omni Technical Report
- VideoLLaMA 3: Frontier Multimodal Foundation Models for Image and Video Understanding
- Meta-Transformer: A Unified Framework for Multimodal Learning
- ChatBridge: Bridging Modalities with Large Language Model as a Language Catalyst
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs