paper-with-me

홈 › Papers

BiLD: Bi-directional Logits Difference Loss for Large Language Model Distillation

2024-06-19 · Minchong Li, Feng Zhou, Xiaohui Song

In recent years, large language models (LLMs) have shown exceptional capabilities across various natural language processing (NLP) tasks. However, such impressive performance often comes with the trade-off of an increased parameter size, posing significant challenges for widespread deployment. Knowledge distillation (KD) provides a solution by transferring knowledge from a large teacher model to a smaller student model. In this paper, we explore the task-specific distillation of LLMs at the logit level. Our investigation reveals that the logits of fine-tuned LLMs exhibit a more extreme long-tail distribution than those from vision models, with hidden "noise" in the long tail affecting distillation performance. Furthermore, existing logits distillation methods often struggle to effectively utilize the internal ranking information from the logits. To address these, we propose the Bi-directional Logits Difference (BiLD) loss. The BiLD loss filters out the long-tail noise by utilizing only top-$k$ teacher and student logits, and leverages the internal logits ranking information by constructing logits differences. To evaluate BiLD loss, we conduct comprehensive experiments on 13 datasets using two types of LLMs. Our results show that the BiLD loss, with only the top-8 logits, outperforms supervised fine-tuning (SFT), vanilla KL loss, and five other distillation methods from both NLP and CV fields.

📄 PDF Abstract BibTeX arXiv:2406.13555

Code (1)

fpcsong/BiLD 공식 구현 pytorch

Tasks

Knowledge DistillationLanguage ModelingLanguage ModellingLarge Language Model

Methods 이 논문이 사용한 방법론

Knowledge Distillation A very simple way to improve the performance of almost any machine learning algorithm is to train many different models on the same data and then to average their predictions.…

Similar Papers 제목 키워드 기반

Switch-KD: Visual-Switch Knowledge Distillation for Vision-Language Models

2026-04-16 · Haoyi Sun, Xiaoxiao Wang, Ning Mao, Qian Wang 외 arxiv

Vision-Language Models (VLMs) have shown remarkable capabilities in joint vision-language understanding, but their large scale poses significant challenges for deployment in resource-constrained scenarios. Knowledge Dist…

Knowledge Distillation

Bidirectional Logits Tree: Pursuing Granularity Reconcilement in Fine-Grained Classification

2024-12-17 · Zhiguang Lu, Qianqian Xu, Shilong Bao, Zhiyong Yang 외

This paper addresses the challenge of Granularity Competition in fine-grained classification tasks, which arises due to the semantic gap between multi-granularity labels. Existing approaches typically develop independent…

A Principled Loss Function for Direct Language Model Alignment

2025-08-10 · Yuandong Tan arxiv

The alignment of large language models (LLMs) with human preferences is commonly achieved through Reinforcement Learning from Human Feedback (RLHF). Direct Preference Optimization (DPO) simplified this paradigm by establ…

Reinforcement Learning

Speculative Decoding with Big Little Decoder

2023-02-15 · NeurIPS 2023 11 · Sehoon Kim, Karttikeya Mangalam, Suhong Moon, Jitendra Malik 외

The recent emergence of Large Language Models based on the Transformer architecture has enabled dramatic advancements in the field of Natural Language Processing. However, these models have long inference latency, which …

Decoderde-enGPUMachine Translation+1

Stabilizing Policy Optimization via Logits Convexity

2026-03-01 · Hongzhan Chen, Tao Yang, Yuhua Zhu, Shiping Gao 외 arxiv

While reinforcement learning (RL) has been central to the recent success of large language models (LLMs), RL optimization is notoriously unstable, especially when compared to supervised fine-tuning (SFT). In this work, w…

Reinforcement Learning