HNHacker News
TopNewBestAskShowJobs

andy12_

480 karma · joined April 1, 2024

submissionscomments

The Obsessed Encoder

enigma.inc·1 pts·andy12_·
0

Nvidia Vera Rubin NVL72 measured to have 10x TPS per megawatt than Blackwell

twitter.com·4 pts·andy12_·
0

Can You Learn to See Without Images? Procedural Warm-Up for Vision Transformers

arxiv.org·2 pts·andy12_·
1

From Memorization to Reasoning in the Spectrum of Loss Curvature

arxiv.org·65 pts·andy12_·
14

Concrete "battery" developed at MIT now packs 10 times the power

news.mit.edu·3 pts·andy12_·
2

Gauss, an Agent for Autoformalization

math.inc·6 pts·andy12_·
0

Spurious Rewards: Rethinking Training Signals in RLVR

rethink-rlvr.notion.site·1 pts·andy12_·
0

VR-CLI: Learning to Reason for Long-Form Story Generation

arxiv.org·2 pts·andy12_·
0

Tokenformer: Rethinking transformer scaling with tokenized model parameters

arxiv.org·3 pts·andy12_·
1

Selective Attention Improves Transformer

arxiv.org·1 pts·andy12_·
1

The AdEMAMix Optimizer: Better, Faster, Older

arxiv.org·2 pts·andy12_·
0