Training Recipe: Strongest Zero-Shot Model Beats General Retrievers
tomaarsen · x · 2026-08-26
The author shares the full recipe for training a strong zero-shot model: using a pre-supervised checkpoint, 1M domain pairs, in-batch negatives, full document length, and a higher-than-usual learning rate. The finetuned model achieves 84.9% rank-1 accuracy, significantly outperforming the zero-shot baseline. Multi-vector/late-interaction architectures are performing exceptionally well.
Related event: Fine-tuning ColBERT on a single RTX 3090 beats general-purpose retrievers(14 posts)→
More from Apps
- ChatGPT desktop shortcuts even an OpenAI employee didn't know about — nickbaumann_ · 2026-10-16
- YC F26's Praxis launches AI agents to run law firm ops end to end — ycombinator · 2026-10-16
- iOS 'dots' workflow: one Siri command turns screenshots into a background shopping and saving agent — nickbaumann_ · 2026-10-16
- Open Nshipyard cleans Canada's messy civic data, exposing $15.1B phantom spend, with MCP tools for agents — RichardsonDx · 2026-10-16
- Ex-DeepMind Eng Lead's AI Calling Agent Fo Cut Hang-Up Rate from 80% to 20-30% — julianweisser · 2026-10-16
- Zero-latency lab-in-the-loop tests whether drug-discovery agents cheat via memorization — niloofar_mire · 2026-10-16