Towards Uncovering How Large Language Model Works: An Explainability Perspective
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866910411678285824 |
|---|---|
| author | Zhao, Haiyan Yang, Fan Shen, Bo Lakkaraju, Himabindu Du, Mengnan |
| author_facet | Zhao, Haiyan Yang, Fan Shen, Bo Lakkaraju, Himabindu Du, Mengnan |
| contents | Large language models (LLMs) have led to breakthroughs in language tasks, yet the internal mechanisms that enable their remarkable generalization and reasoning abilities remain opaque. This lack of transparency presents challenges such as hallucinations, toxicity, and misalignment with human values, hindering the safe and beneficial deployment of LLMs. This paper aims to uncover the mechanisms underlying LLM functionality through the lens of explainability. First, we review how knowledge is architecturally composed within LLMs and encoded in their internal parameters via mechanistic interpretability techniques. Then, we summarize how knowledge is embedded in LLM representations by leveraging probing techniques and representation engineering. Additionally, we investigate the training dynamics through a mechanistic perspective to explain phenomena such as grokking and memorization. Lastly, we explore how the insights gained from these explanations can enhance LLM performance through model editing, improve efficiency through pruning, and better align with human values. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_10688 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Towards Uncovering How Large Language Model Works: An Explainability Perspective Zhao, Haiyan Yang, Fan Shen, Bo Lakkaraju, Himabindu Du, Mengnan Computation and Language Large language models (LLMs) have led to breakthroughs in language tasks, yet the internal mechanisms that enable their remarkable generalization and reasoning abilities remain opaque. This lack of transparency presents challenges such as hallucinations, toxicity, and misalignment with human values, hindering the safe and beneficial deployment of LLMs. This paper aims to uncover the mechanisms underlying LLM functionality through the lens of explainability. First, we review how knowledge is architecturally composed within LLMs and encoded in their internal parameters via mechanistic interpretability techniques. Then, we summarize how knowledge is embedded in LLM representations by leveraging probing techniques and representation engineering. Additionally, we investigate the training dynamics through a mechanistic perspective to explain phenomena such as grokking and memorization. Lastly, we explore how the insights gained from these explanations can enhance LLM performance through model editing, improve efficiency through pruning, and better align with human values. |
| title | Towards Uncovering How Large Language Model Works: An Explainability Perspective |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2402.10688 |