Best of Both Worlds: Fine-Tuned Small Models Complement LLMs for Cost and Performance Benefits
arXiv:2610.24260v1 Announce Type: new Abstract: BERT-based models require thousands of labeled samples to train. Larger models like GPT-X require only 1-5 examples per class. However, performance, cost and time tradeoffs of such models remain under-explored. In this work, we classify survey text da…
Read the full story at arXiv cs.CL ↗
Timeline · 1 report
- 2026-10-21 04:00 · arXiv cs.CL
Best of Both Worlds: Fine-Tuned Small Models Complement LLMs for Cost and Performance Benefits