Cover Image

Where and What Matters: Sensitivity-Aware Task Vectors for Many-Shot Multimodal In-Context Learning

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Ma, Ziyu, Gou, Chenhui, Hu, Yiming, Wang, Yong, Chu, Xiangxiang, Zhuang, Bohan, Cai, Jianfei
Format:	Preprint
Published:	2025
Subjects:	Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Multimodal Task Vectors Enable Many-Shot Multimodal In-Context Learning
by: Huang, Brandon, et al.
Published: (2024)

An Empirical Study on How Video-LLMs Answer Video Questions
by: Gou, Chenhui, et al.
Published: (2025)

FlashBlock: Attention Caching for Efficient Long-Context Block Diffusion
by: Chen, Zhuokun, et al.
Published: (2026)

CoEvolve: Training LLM Agents via Agent-Data Mutual Evolution
by: Yang, Shidong, et al.
Published: (2026)

Robust MAE-Driven NAS: From Mask Reconstruction to Architecture Innovation
by: Hu, Yiming, et al.
Published: (2023)

DrVideo: Document Retrieval Based Long Video Understanding
by: Ma, Ziyu, et al.
Published: (2024)

Efficient Stitchable Task Adaptation
by: He, Haoyu, et al.
Published: (2023)

Many-Shot In-Context Learning
by: Agarwal, Rishabh, et al.
Published: (2024)

Many-Shot In-Context Learning in Multimodal Foundation Models
by: Jiang, Yixing, et al.
Published: (2024)

Selecting Demonstrations for Many-Shot In-Context Learning via Gradient Matching
by: Zhang, Jianfei, et al.
Published: (2025)

D$^2$Evo: Dual Difficulty-Aware Self-Evolution for Data-Efficient Reinforcement Learning
by: Zhang, Ru, et al.
Published: (2026)

What Really Matters in Many-Shot Attacks? An Empirical Study of Long-Context Vulnerabilities in LLMs
by: Kim, Sangyeop, et al.
Published: (2025)

Tree Search for LLM Agent Reinforcement Learning
by: Ji, Yuxiang, et al.
Published: (2025)

SkillClaw: Let Skills Evolve Collectively with Agentic Evolver
by: Ma, Ziyu, et al.
Published: (2026)

Understanding the Prompt Sensitivity
by: Liu, Yang, et al.
Published: (2026)

Compressing Many-Shots in In-Context Learning
by: Khatri, Devvrit, et al.
Published: (2025)

AdapShot: Adaptive Many-Shot In-Context Learning with Semantic-Aware KV Cache Reuse
by: Ou, Jie, et al.
Published: (2026)

Thinking with Map: Reinforced Parallel Map-Augmented Agent for Geolocalization
by: Ji, Yuxiang, et al.
Published: (2026)

On Many-Shot In-Context Learning for Long-Context Evaluation
by: Zou, Kaijian, et al.
Published: (2024)

Evaluating and Advancing Multimodal Large Language Models in Perception Ability Lens
by: Chen, Feng, et al.
Published: (2024)

MARS Policy: Multimodality Only When It Matters
by: Jia, Jindou, et al.
Published: (2026)

GeReA: Question-Aware Prompt Captions for Knowledge-based Visual Question Answering
by: Ma, Ziyu, et al.
Published: (2024)

Label Words as Local Task Vectors in In-Context Learning
by: Zheng, Bowen, et al.
Published: (2024)

Many-Shot In-Context Learning for Molecular Inverse Design
by: Moayedpour, Saeed, et al.
Published: (2024)

Towards Compute-Optimal Many-Shot In-Context Learning
by: Golchin, Shahriar, et al.
Published: (2025)

Sparsity Forcing: Reinforcing Token Sparsity of MLLMs
by: Chen, Feng, et al.
Published: (2025)

Is One-Shot In-Context Learning Helpful for Data Selection in Task-Specific Fine-Tuning of Multimodal LLMs?
by: An, Xiao, et al.
Published: (2026)

Semantic Context Matters: Improving Conditioning for Autoregressive Models
by: Jin, Dongyang, et al.
Published: (2025)

Point-Cache: Test-time Dynamic and Hierarchical Cache for Robust and Generalizable Point Cloud Analysis
by: Sun, Hongyu, et al.
Published: (2025)

Mimic In-Context Learning for Multimodal Tasks
by: Jiang, Yuchu, et al.
Published: (2025)

Distilling Many-Shot In-Context Learning into a Cheat Sheet
by: Honda, Ukyo, et al.
Published: (2025)

MAPLE: Many-Shot Adaptive Pseudo-Labeling for In-Context Learning
by: Chen, Zihan, et al.
Published: (2025)

Adaptive Task Balancing for Visual Instruction Tuning via Inter-Task Contribution and Intra-Task Difficulty
by: Dai, Yanqi, et al.
Published: (2024)

JRDB-PanoTrack: An Open-world Panoptic Segmentation and Tracking Robotic Dataset in Crowded Human Environments
by: Le, Duy-Tho, et al.
Published: (2024)

How Well Can Vision Language Models See Image Details?
by: Gou, Chenhui, et al.
Published: (2024)

QLLM: Accurate and Efficient Low-Bitwidth Quantization for Large Language Models
by: Liu, Jing, et al.
Published: (2023)

ME-Switch: A Memory-Efficient Expert Switching Framework for Large Language Models
by: Liu, Jing, et al.
Published: (2024)

Efficient Many-Shot In-Context Learning with Dynamic Block-Sparse Attention
by: Xiao, Emily, et al.
Published: (2025)

Scaling Laws for Many-Shot In-Context Learning with Self-Generated Annotations
by: Gu, Zhengyao, et al.
Published: (2025)

Label What Matters: Modality-Balanced and Difficulty-Aware Multimodal Active Learning
by: Zeng, Yuqiao, et al.
Published: (2026)