Synthetic bootstrapped pretraining

• URL: http://arxiv.org/abs/2509.15248v2
• Date: Wed, 24 Sep 2025 06:04:40 GMT
• Title: Synthetic bootstrapped pretraining
• Authors: Zitong Yang, Aonan Zhang, Hong Liu, Tatsunori Hashimoto, Emmanuel Candès, Chong Wang, Ruoming Pang,
• Abstract summary: We introduce Synthetic Bootstrapped Pretraining (SBP), a language model (LM) pretraining procedure.<n>SBP first learns a model of relations between documents from the pretraining dataset and then leverages it to synthesize a vast new corpus for joint training.<n>We find SBP consistently improves upon a strong repetition baseline and delivers a significant fraction of performance improvement attainable by an oracle upper bound.
• Score: 52.92577542049469
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We introduce Synthetic Bootstrapped Pretraining (SBP), a language model (LM) pretraining procedure that first learns a model of relations between documents from the pretraining dataset and then leverages it to synthesize a vast new corpus for joint training. While the standard pretraining teaches LMs to learn causal correlations among tokens within a single document, it is not designed to efficiently model the rich, learnable inter-document correlations that can potentially lead to better performance. We validate SBP by designing a compute-matched pretraining setup and pretrain a 3B-parameter model on up to 1T tokens from scratch. We find SBP consistently improves upon a strong repetition baseline and delivers a significant fraction of performance improvement attainable by an oracle upper bound with access to 20x more unique data. Qualitative analysis reveals that the synthesized documents go beyond mere paraphrases -- SBP first abstracts a core concept from the seed material and then crafts a new narration on top of it. Besides strong empirical performance, SBP admits a natural Bayesian interpretation: the synthesizer implicitly learns to abstract the latent concepts shared between related documents.

Related papers

• Paired by the Teacher: Turning Unpaired Data into High-Fidelity Pairs for Low-Resource Text Generation [17.879274676067784]
  Paired by the Teacher (PbT) is a two-stage teacher-student pipeline that synthesizes accurate input-output pairs without human labels or parallel data.<n>We evaluate PbT on five benchmarks-document summarization, dialogue summarization, SAMSum, DialogSum, and question generation.<n>An 8B student trained only on PbT data outperforms models trained on 70 B teacher-generated corpora.
  arXiv  Detail & Related papers  (2025-09-29T17:51:55Z)
• Learning and Transferring Sparse Contextual Bigrams with Linear Transformers [47.37256334633102]
  We introduce the Sparse Con Bigram model, where the next token's generation depends on a sparse set of earlier positions determined by the last token. We analyze the training dynamics and sample complexity of learning SCB using a one-layer linear transformer with a gradient-based algorithm. We prove that, provided a nontrivial correlation between the downstream and pretraining tasks, finetuning from a pretrained model allows us to bypass the initial sample-intensive stage.
  arXiv  Detail & Related papers  (2024-10-30T20:29:10Z)
• Synthetic continued pretraining [29.6872772403251]
  We propose synthetic continued pretraining on a small corpus of domain-specific documents. We instantiate this proposal with EntiGraph, a synthetic data augmentation algorithm. We show how synthetic data augmentation can "rearrange" knowledge to enable more data-efficient learning.
  arXiv  Detail & Related papers  (2024-09-11T17:21:59Z)
• The Fine Line: Navigating Large Language Model Pretraining with Down-streaming Capability Analysis [27.310894780313618]
  This paper undertakes a comprehensive comparison of model capabilities at various pretraining intermediate checkpoints. We confirm that specific downstream metrics exhibit similar training dynamics across models of different sizes. In addition to our core findings, we've reproduced Amber and OpenLLaMA, releasing their intermediate checkpoints.
  arXiv  Detail & Related papers  (2024-04-01T16:00:01Z)
• Document-Level In-Context Few-Shot Relation Extraction via Pre-Trained Language Models [29.94694305204144]
  We present a novel framework for document-level in-context few-shot relation extraction. We evaluate our framework using DocRED, the largest publicly available dataset for document-level relation extraction.
  arXiv  Detail & Related papers  (2023-10-17T09:10:27Z)
• Scalable Learning of Latent Language Structure With Logical Offline Cycle Consistency [71.42261918225773]
  Conceptually, LOCCO can be viewed as a form of self-learning where the semantic being trained is used to generate annotations for unlabeled text. As an added bonus, the annotations produced by LOCCO can be trivially repurposed to train a neural text generation model.
  arXiv  Detail & Related papers  (2023-05-31T16:47:20Z)
• Synthetic Pre-Training Tasks for Neural Machine Translation [16.6378815054841]
  Our goal is to understand the factors that contribute to the effectiveness of pre-training models when using synthetic resources. We propose several novel approaches to pre-training translation models that involve different levels of lexical and structural knowledge. Our experiments on multiple language pairs reveal that pre-training benefits can be realized even with high levels of obfuscation or purely synthetic parallel data.
  arXiv  Detail & Related papers  (2022-12-19T21:34:00Z)
• Unifying Language Learning Paradigms [96.35981503087567]
  We present a unified framework for pre-training models that are universally effective across datasets and setups. We show how different pre-training objectives can be cast as one another and how interpolating between different objectives can be effective. Our model also achieve strong results at in-context learning, outperforming 175B GPT-3 on zero-shot SuperGLUE and tripling the performance of T5-XXL on one-shot summarization.
  arXiv  Detail & Related papers  (2022-05-10T19:32:20Z)
• An Explanation of In-context Learning as Implicit Bayesian Inference [117.19809377740188]
  We study the role of the pretraining distribution on the emergence of in-context learning. We prove that in-context learning occurs implicitly via Bayesian inference of the latent concept. We empirically find that scaling model size improves in-context accuracy even when the pretraining loss is the same.
  arXiv  Detail & Related papers  (2021-11-03T09:12:33Z)
• Syntactic Structure Distillation Pretraining For Bidirectional Encoders [49.483357228441434]
  We introduce a knowledge distillation strategy for injecting syntactic biases into BERT pretraining. We distill the approximate marginal distribution over words in context from the syntactic LM. Our findings demonstrate the benefits of syntactic biases, even in representation learners that exploit large amounts of data.
  arXiv  Detail & Related papers  (2020-05-27T16:44:01Z)
• Pre-training for Abstractive Document Summarization by Reinstating Source Text [105.77348528847337]
  This paper presents three pre-training objectives which allow us to pre-train a Seq2Seq based abstractive summarization model on unlabeled text. Experiments on two benchmark summarization datasets show that all three objectives can improve performance upon baselines.
  arXiv  Detail & Related papers  (2020-04-04T05:06:26Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.