Reference library

Language Models are Few-Shot Learners

Paper summary

Evaluates GPT-3 on tasks supplied through instructions and a few examples without task-specific gradient updates during evaluation. Scaling pretraining yields substantial in-context task performance while leaving uneven capabilities and limitations.

Relevance to the project

Clarifies the difference between context-dependent behavior and lasting parameter adaptation. Useful when cataloging what an ordinary running LLM does and lacks.

Language modelsLanguage model trainingScaling lawsLearning and adaptation

Few-shot performance does not establish autonomous maintenance or persistent lifetime learning.

Source record

Reference ID: arxiv-2005-14165
Archived version: Publisher/institution PDF advertised by article metadata; 2005.14165v4
Summary provenance: reference-library/papers/arxiv-2005-14165/metadata.json · summary
Review status: initial orientation summary; not a critical full-paper review
Reading scope: Selected abstract/opening and available closing discussion/conclusion passages in the preserved text extract; broader project notes provide context. No reproduction of experiments, exhaustive methods review, or equation/figure audit.