Pith. sign in

REVIEW 4 cited by

Pretrained Transformers as Universal Computation Engines

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2103.05247 v2 pith:63NO5ZVK submitted 2021-03-09 cs.LG cs.AI

classification cs.LGcs.AI
keywords finetuningperformancepretrainedtaskstransformercomputationinvestigatelanguage
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

We investigate the capability of a transformer pretrained on natural language to generalize to other modalities with minimal finetuning -- in particular, without finetuning of the self-attention and feedforward layers of the residual blocks. We consider such a model, which we call a Frozen Pretrained Transformer (FPT), and study finetuning it on a variety of sequence classification tasks spanning numerical computation, vision, and protein fold prediction. In contrast to prior works which investigate finetuning on the same modality as the pretraining dataset, we show that pretraining on natural language can improve performance and compute efficiency on non-language downstream tasks. Additionally, we perform an analysis of the architecture, comparing the performance of a random initialized transformer to a random LSTM. Combining the two insights, we find language-pretrained transformers can obtain strong performance on a variety of non-language tasks.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 4 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Random Initialization Can't Catch Up: The Advantage of Language Model Transfer for Time Series Forecasting

    cs.CL 2025-06 conditional novelty 5.0 of 10

    Pretrained T5 language weights give a persistent validation-loss advantage over random initialization for low-data time series forecasting, and the advantage does not vanish within the training budget.

  2. LLM4SG: Adapting Large Language Model for Scatterer Generation via Synesthesia of Machines

    eess.SP 2025-05 conditional novelty 5.0 of 10

    Fine-tuning a small GPT-2 with LoRA on a new synthetic V2V dataset lets it predict ray-tracing scatterer grids from LiDAR point clouds, outperforming a ResNet baseline.

  3. Transfer of Structural Knowledge from Synthetic Languages

    cs.CL 2025-05 conditional novelty 5.0 of 10

    A new synthetic language, flat_shuffle, transfers more structure to English fine-tuning than earlier synthetic bracket languages, though still far short of training on English from scratch.

  4. A Vision Toward Energy-Efficient Domain-Specific Artificial Intelligence Models and Agents

    cs.AI 2025-10 unverdicted novelty 2.0 of 10

    A position paper proposing compact, domain-specific AI agents as the path to ≥1000× energy efficiency, without demonstrating the claim.

Pith tools