paper-with-me

홈 › Papers

AI capabilities can be significantly improved without expensive retraining

2023-12-12 · Tom Davidson, Jean-Stanislas Denain, Pablo Villalobos, Guillem Bas

State-of-the-art AI systems can be significantly improved without expensive retraining via "post-training enhancements"-techniques applied after initial training like fine-tuning the system to use a web browser. We review recent post-training enhancements, categorizing them into five types: tool-use, prompting methods, scaffolding, solution selection, and data generation. Different enhancements improve performance on different tasks, making it hard to compare their significance. So we translate improvements from different enhancements into a common currency, the compute-equivalent gain: how much additional training compute would be needed to improve performance by the same amount as the enhancement. Our non-experimental work shows that post-training enhancements have significant benefits: most surveyed enhancements improve benchmark performance by more than a 5x increase in training compute, some by more than 20x. Post-training enhancements are relatively cheap to develop: fine-tuning costs are typically <1% of the original training cost. Governing the development of capable post-training enhancements may be challenging because frontier models could be enhanced by a wide range of actors.

📄 PDF Abstract BibTeX arXiv:2312.07413

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Train Separately, Merge Together: Modular Post-Training with Mixture-of-Experts

2026-04-20 · Jacob Morrison, Sanjay Adhikesaven, Akshita Bhagia, Matei Zaharia 외 arxiv

Extending a fully post-trained language model with new domain capabilities is fundamentally limited by monolithic training paradigms: retraining from scratch is expensive and scales poorly, while continued training often…

Reinforcement Learning

Extending the Context of Pretrained LLMs by Dropping Their Positional Embeddings

2025-12-13 · Yoav Gelberg, Koshi Eguchi, Takuya Akiba, Edoardo Cetin arxiv

So far, expensive finetuning beyond the pretraining sequence length has been a requirement for effectively extending the context of language models (LM). In this work, we break this key bottleneck by Dropping the Positio…

Revisiting Pretraining with Adapters

2021-08-01 · ACL (RepL4NLP) 2021 8 · Seungwon Kim, Alex Shum, Nathan Susanj, Jonathan Hilgart

Pretrained language models have served as the backbone for many state-of-the-art NLP results. These models are large and expensive to train. Recent work suggests that continued pretraining on task-specific data is worth …

Continual PretrainingTransfer Learning

Are Language Models Up to Sequential Optimization Problems? From Evaluation to a Hegelian-Inspired Enhancement

2025-02-04 · Soheil Abbasloo

Large Language Models (LLMs) have demonstrated impressive capabilities across numerous fields, presenting an opportunity to revolutionize optimization problem-solving, a crucial, ubiquitous, and complex domain. This pape…

Learning without Forgetting

2016-06-29 · Zhizhong Li, Derek Hoiem

When building a unified vision system or gradually adding new capabilities to a system, the usual assumption is that training data for all tasks is always available. However, as the number of tasks grows, storing and ret…

Class Incremental LearningContinual LearningDisjoint 10-1Disjoint 15-1+8