Text Capability Loss in Vision-Language Adaptation: An Attention-Sink Diagnosis
cs.LG
Submitted: 2026-09-01
Updated: 2026-09-01
Code: https://github.com/EQ-bench/EQ-Bench
Terminology
Sources
- Program Synthesis with Large Language Models
- Qwen3-VL Technical Report
- Qwen2.5-VL Technical Report
- Training Verifiers to Solve Math Word Problems
- The Llama 3 Herd of Models
- Mistral 7B
- SEED-Bench: Benchmarking Multimodal LLMs with Generative Comprehension
- Ovis: Structural Embedding Alignment for Multimodal Large Language Model
- InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency
- In-context Learning and Induction Heads
- EQ-Bench: An Emotional Intelligence Benchmark for Large Language Models
- Olmo 3
- Qwen3 Technical Report
- Qwen2 Technical Report
- Qwen2.5 Technical Report
- MiniCPM-V 4.5: Cooking Efficient MLLMs via Architecture, Data, and Training Recipe
- Instruction-Following Evaluation for Large Language Models
- InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Models
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks