Rosinality(@rosinality):https://t.co/uJ3YSeOlhE Recurrent transformer that injects activations from the top layers to the bottom layers at the next step (https://t.co/bubRFyPaBt). Why does this work without training?

注册并分享邀请链接，可获得视频播放与邀请奖励。

立即注册

Rosinality

@rosinality

ML Engineer @poolsideai

加入 October 2008

1K 正在关注 8K 粉丝

Rosinality@rosinality

2026.08.19 08:11

Recurrent transformer that injects activations from the top layers to the bottom layers at the next step ( Why does this work without training?

显示更多

0

0

32

1.1K

105

转发到社区

热门用户

0 粉丝

4.2M 粉丝

569K 粉丝

26.4M 粉丝

オリコンニュース

1.9M 粉丝

吴说区块链

@wublockchain12

181.5K 粉丝

@Kongkongda5882

34.1K 粉丝

100.9K 粉丝

ツイッター速報〜BreakingNews

256.8K 粉丝

0 粉丝

中國新聞社

547.4K 粉丝

348.8K 粉丝

华尔街观察 Xtrader

@cnfinancewatch

129.7K 粉丝

23.7K 粉丝

3.4M 粉丝