The representation landscape of few-shot learning and fine-tuning in large language models
In-context learning (ICL) and supervised fine-tuning (SFT) are two common strategies for improving the performance of modern large language models (LLMs) on specific tasks. Despite their different natures, these strategies often lead to comparable performance gains. However, little is known about whether they induce similar representations inside LLMs. We approach this problem by analyzing the probability landscape of their hidden representations in the two cases. More specifically, we compare how LLMs solve the same question-answering task, finding that ICL and SFT create very different internal structures, in both cases undergoing a sharp transition in the middle of the network. In the first half of the network, ICL shapes interpretable representations hierarchically organized according to their semantic content. In contrast, the probability landscape obtained with SFT is fuzzier and semantically mixed. In the second half of the model, the fine-tuned representations develop probability modes that better encode the identity of answers, while the landscape of ICL representations is characterized by less defined peaks. Our approach reveals the diverse computational strategies developed inside LLMs to solve the same task across different conditions, allowing us to make a step towards designing optimal methods to extract information from language models.
Code (1)
Tasks
Few-Shot LearningIn-Context LearningQuestion AnsweringMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
Flatten Long-Range Loss Landscapes for Cross-Domain Few-Shot Learning
Cross-domain few-shot learning (CDFSL) aims to acquire knowledge from limited training data in the target domain by leveraging prior knowledge transferred from source domains with abundant training samples. CDFSL faces c…
Cross-Domain Few-Shotcross-domain few-shot learningFew-Shot LearningNavigating the Landscape of Large Language Models: A Comprehensive Review and Analysis of Paradigms and Fine-Tuning Strategies
With the surge of ChatGPT,the use of large models has significantly increased,rapidly rising to prominence across the industry and sweeping across the internet. This article is a comprehensive review of fine-tuning metho…
Few-Shot LearningKnowledge DistillationMulti-Task Learningparameter-efficient fine-tuningTransfer Learning for Finetuning Large Language Models
As the landscape of large language models expands, efficiently finetuning for specific tasks becomes increasingly crucial. At the same time, the landscape of parameter-efficient finetuning methods rapidly expands. Conseq…
Bayesian OptimizationMeta-LearningTransfer LearningExploring The Landscape of Distributional Robustness for Question Answering Models
We conduct a large empirical evaluation to investigate the landscape of distributional robustness in question answering. Our investigation spans over 350 models and 16 question answering datasets, including a diverse set…
In-Context LearningQuestion AnsweringDelta Activations: A Representation for Finetuned Large Language Models
The success of powerful open source Large Language Models (LLMs) has enabled the community to create a vast collection of post-trained models adapted to specific tasks and domains. However, navigating and understanding t…