Pearl: Personalizing Large Language Model Writing Assistants with Generation-Calibrated Retrievers
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2023
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910684098330624 |
|---|---|
| author | Mysore, Sheshera Lu, Zhuoran Wan, Mengting Yang, Longqi Sarrafzadeh, Bahareh Menezes, Steve Baghaee, Tina Gonzalez, Emmanuel Barajas Neville, Jennifer Safavi, Tara |
| author_facet | Mysore, Sheshera Lu, Zhuoran Wan, Mengting Yang, Longqi Sarrafzadeh, Bahareh Menezes, Steve Baghaee, Tina Gonzalez, Emmanuel Barajas Neville, Jennifer Safavi, Tara |
| contents | Powerful large language models have facilitated the development of writing assistants that promise to significantly improve the quality and efficiency of composition and communication. However, a barrier to effective assistance is the lack of personalization in LLM outputs to the author's communication style, specialized knowledge, and values. In this paper, we address this challenge by proposing Pearl, a LLM writing assistant personalized with a retriever that is trained to be generation-calibrated for personalization. Generation calibration ensures that our retriever selects historic user authored documents to augment an LLM prompt such that they are likely to help an LLM generation better adhere to a users' preferences. We propose two key novelties for training such a retriever: (1) A training data selection method that identifies user requests likely to benefit from personalization and documents that provide that benefit; and (2) A scale-calibrating KL-divergence objective that ensures that our retriever scores remain proportional to the downstream generation quality from using the document for personalized generation. In a series of holistic evaluations, we demonstrate the effectiveness of Pearl in generating long-form texts on multiple social media datasets. Finally, we demonstrate how a generation-calibrated retriever can double as a performance predictor -- detecting low quality retrieval, and improving potentially under-performing outputs via revision with LLMs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2311_09180 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Pearl: Personalizing Large Language Model Writing Assistants with Generation-Calibrated Retrievers Mysore, Sheshera Lu, Zhuoran Wan, Mengting Yang, Longqi Sarrafzadeh, Bahareh Menezes, Steve Baghaee, Tina Gonzalez, Emmanuel Barajas Neville, Jennifer Safavi, Tara Computation and Language Human-Computer Interaction Information Retrieval Powerful large language models have facilitated the development of writing assistants that promise to significantly improve the quality and efficiency of composition and communication. However, a barrier to effective assistance is the lack of personalization in LLM outputs to the author's communication style, specialized knowledge, and values. In this paper, we address this challenge by proposing Pearl, a LLM writing assistant personalized with a retriever that is trained to be generation-calibrated for personalization. Generation calibration ensures that our retriever selects historic user authored documents to augment an LLM prompt such that they are likely to help an LLM generation better adhere to a users' preferences. We propose two key novelties for training such a retriever: (1) A training data selection method that identifies user requests likely to benefit from personalization and documents that provide that benefit; and (2) A scale-calibrating KL-divergence objective that ensures that our retriever scores remain proportional to the downstream generation quality from using the document for personalized generation. In a series of holistic evaluations, we demonstrate the effectiveness of Pearl in generating long-form texts on multiple social media datasets. Finally, we demonstrate how a generation-calibrated retriever can double as a performance predictor -- detecting low quality retrieval, and improving potentially under-performing outputs via revision with LLMs. |
| title | Pearl: Personalizing Large Language Model Writing Assistants with Generation-Calibrated Retrievers |
| topic | Computation and Language Human-Computer Interaction Information Retrieval |
| url | https://arxiv.org/abs/2311.09180 |