Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2410.23463 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866912351926616064 |
|---|---|
| author | Liu, Gabrielle Kaili-May Shi, Bowen Caciularu, Avi Szpektor, Idan Cohan, Arman |
| author_facet | Liu, Gabrielle Kaili-May Shi, Bowen Caciularu, Avi Szpektor, Idan Cohan, Arman |
| contents | Multi-document (MD) processing is crucial for LLMs to handle real-world tasks such as summarization and question-answering across large sets of documents. While LLMs have improved at processing long inputs, MD contexts still present unique difficulties, including management of inter-document dependencies, redundancy, and incoherent structures. To address this challenge, we introduce MDCure, a scalable and effective instruction data generation framework to enhance the MD capabilities of LLMs without the computational cost of pre-training or reliance on human-annotated data. MDCure generates high-quality synthetic MD instruction data over sets of articles via targeted prompts. We also introduce MDCureRM, a cost-effective, MD-specific reward model to score and filter generated data based on their training utility for MD settings. MDCure is compatible with open- and closed-source models in addition to policy optimization methods such as PPO, enabling even small open-source models to surpass proprietary LLMs as strong generators of high-quality MD instruction data without further data filtering. With MDCure, we fine-tune a wide variety of LLMs up to 70B parameters in size from the FlanT5, Qwen2, and LLAMA3.1 model families. Extensive evaluations on a wide range of MD and long-context benchmarks spanning various tasks and domains show MDCure consistently improves performance over pre-trained baselines and base models by up to 75.1%. Our code, datasets, and models are available at https://github.com/yale-nlp/MDCure. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_23463 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | MDCure: A Scalable Pipeline for Multi-Document Instruction-Following Liu, Gabrielle Kaili-May Shi, Bowen Caciularu, Avi Szpektor, Idan Cohan, Arman Computation and Language Machine Learning Multi-document (MD) processing is crucial for LLMs to handle real-world tasks such as summarization and question-answering across large sets of documents. While LLMs have improved at processing long inputs, MD contexts still present unique difficulties, including management of inter-document dependencies, redundancy, and incoherent structures. To address this challenge, we introduce MDCure, a scalable and effective instruction data generation framework to enhance the MD capabilities of LLMs without the computational cost of pre-training or reliance on human-annotated data. MDCure generates high-quality synthetic MD instruction data over sets of articles via targeted prompts. We also introduce MDCureRM, a cost-effective, MD-specific reward model to score and filter generated data based on their training utility for MD settings. MDCure is compatible with open- and closed-source models in addition to policy optimization methods such as PPO, enabling even small open-source models to surpass proprietary LLMs as strong generators of high-quality MD instruction data without further data filtering. With MDCure, we fine-tune a wide variety of LLMs up to 70B parameters in size from the FlanT5, Qwen2, and LLAMA3.1 model families. Extensive evaluations on a wide range of MD and long-context benchmarks spanning various tasks and domains show MDCure consistently improves performance over pre-trained baselines and base models by up to 75.1%. Our code, datasets, and models are available at https://github.com/yale-nlp/MDCure. |
| title | MDCure: A Scalable Pipeline for Multi-Document Instruction-Following |
| topic | Computation and Language Machine Learning |
| url | https://arxiv.org/abs/2410.23463 |