7 papers
PruneCD: Contrasting Pruned Self Model to Improve Decoding Factuality
Byeongho Yu, Changhun Lee, Jungyu Jin +1
To mitigate the hallucination problem in large language models, DoLa exploits early exit logits from the same model as a contrastive prior. However, we found that these early exit…
AMQ: Enabling AutoML for Mixed-precision Weight-Only Quantization of Large Language Models
Sangjun Lee, Seung-taek Woo, Jungyu Jin +2
To enable broader deployment of Large Language Models (LLMs), it is essential to identify the best-performing model under strict memory constraints. We present AMQ, Automated Mixed…
Merge-Friendly Post-Training Quantization for Multi-Target Domain Adaptation
Juncheol Shin, Minsang Seok, Seonggon Kim +1
Model merging has emerged as a powerful technique for combining task-specific weights, achieving superior performance in multi-target domain adaptation. However, when applied to pr…
GraLoRA: Granular Low-Rank Adaptation for Parameter-Efficient Fine-Tuning
Yeonjoon Jung, Daehyun Ahn, Hyungjun Kim +2
Low-Rank Adaptation (LoRA) is a popular method for parameter-efficient fine-tuning (PEFT) of generative models, valued for its simplicity and effectiveness. Despite recent enhancem…
HOT: Hadamard-based Optimized Training
Seonggon Kim, Juncheol Shin, Seung-taek Woo +1
It has become increasingly important to optimize backpropagation to reduce memory usage and computational overhead. Achieving this goal is highly challenging, as multiple objective…
SEAL: Scaling to Emphasize Attention for Long-Context Retrieval
Changhun Lee, Minsang Seok, Jun-gyu Jin +2
While many advanced LLMs are designed to handle long sequence data, we can still observe notable quality degradation even within the sequence limit. In this work, we introduce a no…