Customizing Large Language Model Generation Style using Parameter-Efficient Finetuning
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866910593232928768 |
|---|---|
| author | Liu, Xinyue Diddee, Harshita Ippolito, Daphne |
| author_facet | Liu, Xinyue Diddee, Harshita Ippolito, Daphne |
| contents | One-size-fits-all large language models (LLMs) are increasingly being used to help people with their writing. However, the style these models are trained to write in may not suit all users or use cases. LLMs would be more useful as writing assistants if their idiolect could be customized to match each user. In this paper, we explore whether parameter-efficient finetuning (PEFT) with Low-Rank Adaptation can effectively guide the style of LLM generations. We use this method to customize LLaMA-2 to ten different authors and show that the generated text has lexical, syntactic, and surface alignment with the target author but struggles with content memorization. Our findings highlight the potential of PEFT to support efficient, user-level customization of LLMs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_04574 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Customizing Large Language Model Generation Style using Parameter-Efficient Finetuning Liu, Xinyue Diddee, Harshita Ippolito, Daphne Computation and Language One-size-fits-all large language models (LLMs) are increasingly being used to help people with their writing. However, the style these models are trained to write in may not suit all users or use cases. LLMs would be more useful as writing assistants if their idiolect could be customized to match each user. In this paper, we explore whether parameter-efficient finetuning (PEFT) with Low-Rank Adaptation can effectively guide the style of LLM generations. We use this method to customize LLaMA-2 to ten different authors and show that the generated text has lexical, syntactic, and surface alignment with the target author but struggles with content memorization. Our findings highlight the potential of PEFT to support efficient, user-level customization of LLMs. |
| title | Customizing Large Language Model Generation Style using Parameter-Efficient Finetuning |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2409.04574 |