Unverified50% confidenceFactExact time

DeepMind's previous Enformer model has an effective input window of approximately 200 kb.

1

Sources

50%

Confidence

Long-term

Relevance

9/10/2026

First Seen

Sources

AlphaGenome Atlas Explained: How DeepMind Uses AI to Decode 3 Billion Base Pairs

Related Entities

DeepMindcompany Enformermodel

Related Claims

UnverifiedDeepSeek-Coder具备16K甚至更长的上下文窗口70% similar UnverifiedGoogle DeepMind released Gemma 4 quantization-aware training weights, reducing the memory footprint of the 12B parameter model to approximately 1GB through targeted 2-bit compression technology.70% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/887348

API

curl https://kongchang.com/api/v1/knowledge/claims/887348

MCP

get_claim(id=887348)