Comment Comment by throwa356262 The tweet mentioned deepseek V4 flash.From HF: 284B parameters (13B active), 1M context window.This is indeed some kind of compressed context and the quality goes down as the context grows. IIRC the V4 paper had some numbers on thishttps://huggingface.co/deepseek-ai/DeepSeek-V4-Flash parent throwa356262 5 months ago
From HF: 284B parameters (13B active), 1M context window.
This is indeed some kind of compressed context and the quality goes down as the context grows. IIRC the V4 paper had some numbers on this
https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash