gpt3-finetune-mnli-40-6-to-89-5
IN premise — summaries/2026/08/24/hu-2021-lora-sR-references-chunk-1.md
Created 2026-08-24T17:10:55+00:00
On GPT-3, fine-tuning improves MNLI accuracy from 40.6% (few-shot prompting) to 89.5% and RTE from 69.0% to 85.4%, demonstrating parameter updates remain essential even at 175B scale
Summary
Even a 175-billion-parameter model like GPT-3 can't reliably do natural language inference just from prompts; actually updating its weights through fine-tuning is what pushes accuracy from chance-level to strong. In practical terms, this means scale alone doesn't replace training, and any system relying on GPT-3 for these tasks still needs the overhead of a fine-tuning step to get usable performance.