From Drafts to Answers: Unlocking LLM Potential via Aggregation Fine-Tuning

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Li, Yafu, Wang, Zhilin, Fu, Tingchen, Cui, Ganqu, Yang, Sen, Cheng, Yu
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866929683212271616
author Li, Yafu
Wang, Zhilin
Fu, Tingchen
Cui, Ganqu
Yang, Sen
Cheng, Yu
author_facet Li, Yafu
Wang, Zhilin
Fu, Tingchen
Cui, Ganqu
Yang, Sen
Cheng, Yu
contents Scaling data and model size has been proven effective for boosting the performance of large language models. In addition to training-time scaling, recent studies have revealed that increasing test-time computational resources can further improve performance. In this work, we introduce Aggregation Fine-Tuning (AFT), a supervised finetuning paradigm where the model learns to synthesize multiple draft responses, referred to as proposals, into a single, refined answer, termed aggregation. At inference time, a propose-and-aggregate strategy further boosts performance by iteratively generating proposals and aggregating them. Empirical evaluations on benchmark datasets show that AFT-trained models substantially outperform standard SFT. Notably, an AFT model, fine-tuned from Llama3.1-8B-Base with only 64k data, achieves a 41.3% LC win rate on AlpacaEval 2, surpassing significantly larger LLMs such as Llama3.1-405B-Instruct and GPT4. By combining sequential refinement and parallel sampling, the propose-and-aggregate framework scales inference-time computation in a flexible manner. Overall, These findings position AFT as a promising approach to unlocking additional capabilities of LLMs without resorting to increasing data volume or model size.
format Preprint
id arxiv_https___arxiv_org_abs_2501_11877
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle From Drafts to Answers: Unlocking LLM Potential via Aggregation Fine-Tuning
Li, Yafu
Wang, Zhilin
Fu, Tingchen
Cui, Ganqu
Yang, Sen
Cheng, Yu
Computation and Language
Artificial Intelligence
Scaling data and model size has been proven effective for boosting the performance of large language models. In addition to training-time scaling, recent studies have revealed that increasing test-time computational resources can further improve performance. In this work, we introduce Aggregation Fine-Tuning (AFT), a supervised finetuning paradigm where the model learns to synthesize multiple draft responses, referred to as proposals, into a single, refined answer, termed aggregation. At inference time, a propose-and-aggregate strategy further boosts performance by iteratively generating proposals and aggregating them. Empirical evaluations on benchmark datasets show that AFT-trained models substantially outperform standard SFT. Notably, an AFT model, fine-tuned from Llama3.1-8B-Base with only 64k data, achieves a 41.3% LC win rate on AlpacaEval 2, surpassing significantly larger LLMs such as Llama3.1-405B-Instruct and GPT4. By combining sequential refinement and parallel sampling, the propose-and-aggregate framework scales inference-time computation in a flexible manner. Overall, These findings position AFT as a promising approach to unlocking additional capabilities of LLMs without resorting to increasing data volume or model size.
title From Drafts to Answers: Unlocking LLM Potential via Aggregation Fine-Tuning
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2501.11877