CRAM proposes a novel framework for Multimodal Continual Instruction Tuning that balances task isolation and parameter efficiency by using centroid-guided routing and adaptive MoE to prevent catastrop…

The paper introduces RA-MoE, a novel fine-tuning framework that leverages the internal routing structure of Mixture-of-Experts (MoE) models to improve performance on multilingual downstream tasks by a…

The paper introduces and evaluates five parameter alignment strategies that significantly mitigate catastrophic forgetting when continually pretraining multilingual expert language models across multi…

The paper introduces Andes, a framework that treats data generation as a plug-and-play agent skill, enabling autonomous alignment of LLMs by providing an intelligent, closed-loop data synthesis interf…

The paper proposes VLM3, a simple, scalable method that demonstrates standard Vision Language Models (VLMs) can natively learn 3D understanding by focusing on architectural simplicity and specific dat…

ConMoE proposes a train-free method for compressing Mixture-of-Experts (MoE) models by consolidating the large expert pool into a smaller set of reusable prototypes and deterministically remapping all…

The paper proposes Dynamic Adapter Routing (DAR), a novel method that significantly improves continual multimodal retrieval by adaptively selecting and merging specialized adapters.

The paper proposes AlignG, a method that learns context-conditioned predicate semantics by using prototype feedback to adapt relation representations based on image-specific evidence, significantly im…

The paper introduces State-Adaptive Prompt Optimization (SAPO), a novel training strategy that treats prompts as dynamic variables to achieve robust fine-tuning, significantly mitigating catastrophic…

The paper introduces DOMINO, a novel inductive framework that synthesizes domain-specific data for LLMs using only reference examples, significantly improving performance on challenging, implicitly de…

VLA-Pro is a plug-and-play framework that enhances cross-task generalization in Vision-Language-Action models by storing and dynamically retrieving task-specific procedural memories, achieving signifi…

Prompt Codebooks (PCO) introduces a compositional framework that treats prompt optimization as discrete learning over reusable instruction units, significantly improving LLM performance while drastica…

The paper proposes SubFit, a novel compression technique that achieves superior LLM compression by replacing non-contiguous, submodule-level components (Attention and FeedForward) with lightweight res…

This paper introduces Anchored Weight Decay (AWD), a regularization technique that effectively prevents prior-task forgetting during LLM fine-tuning with Evolution Strategies (ES), positioning ES as a…

MASER is a lightweight framework that dynamically routes a shared Vision-Language Model (VLM) to the most appropriate modality-specific adapter (e.g., point cloud, RGB) based on the input question, si…

TRACE proposes a novel method to mitigate catastrophic forgetting in continual LLM fine-tuning by identifying and isolating a small, task-specific subset of essential parameters for each task.

The paper introduces CV-Arena, a large-scale open benchmark for instructional computer vision, demonstrating that professional-grade image editing requires advanced capabilities in physical reasoning…

The paper introduces Parallax, a scalable and numerically stable parameterized Local Linear Attention mechanism that significantly improves LLM performance and efficiency compared to existing methods…