Edge Rewrite
// HTMLRewriter · presentation

This page was redesigned at the edge.

Cloudflare fetched the original article and streamed it through HTMLRewriter to apply an entirely new visual system without rebuilding the source page.

Jump to content

Talk:Evo (AI)

Page contents not supported in other languages.
Add topic
From Wikipedia, the free encyclopedia
Latest comment: 23 hours ago by Andrew Davidson in topic Zero-shot

Zero-shot

[edit]

It expanded the model context window to a million base pairs (1 megabase) at single-nucleotide resolution, allowing zero-shot effect prediction for human non-coding variants, BRCA1 disease mutations, and chromatin accessibility patterns.

Zero-shot effect prediction? Can this be clarified? From my stats (but no bio) background that seems like it's supposed to mean "they predicted this gene would cause breast cancer despite providing no data on breast cancer", which seems extremely unlikely. – Closed Limelike Curves (talk) 02:50, 12 August 2026 (UTC)Reply

As I understand it, zero-shot means that it was able to predict the effect of sequences that it had not been specifically trained on. This Nature paper goes into detail: Genome modelling and design across all domains of life with Evo 2. And here's a step-by-step tutorial: Zero-Shot BRCA1 Variant Prediction with Evo 2: A Tutorial. Andrew🐉(talk) 16:41, 12 August 2026 (UTC)Reply