zagwdt

3 ms·

1.
Score Centering Stabilizes Off-Policy Reinforcement Learning(arxiv.org)
3 points by zagwdt 1mo ago|0 comments
2.
A/B testing LLMs in production(together.ai)
2 points by zagwdt 2mo ago|0 comments
3.
Inference Optimization for MiniMax Sparse Attention(together.ai)
1 points by zagwdt 5mo ago|0 comments
4.
by zagwdt 6mo ago
ngl lost alot of trust in cc after reading this, specially point 1 how do you just do that to millions of users building prod code with your shit
5.
DeepSeek V4 in vLLM: Efficient Long-Context Attention(vllm-website-pdzeaspbm-inferact-inc.vercel.app)
3 points by zagwdt 6mo ago|0 comments
6.
Introspective Diffusion Language Models(introspective-diffusion.github.io)
281 points by zagwdt 6mo ago|55 comments
7.
EinsteinArena: Harnessing the collective intelligence of agents in the wild(einsteinarena.com)
5 points by zagwdt 6mo ago|0 comments
8.
RL Meets Adaptive Speculative Training(together.ai)
2 points by zagwdt 7mo ago|0 comments
9.
Weak models excel at long context tasks(together.ai)
2 points by zagwdt 7mo ago|0 comments
10.
TorchSpec: Speculative Decoding Training at Scale(pytorch.org)
2 points by zagwdt 7mo ago|0 comments
11.
Flash Attention 4(together.ai)
1 points by zagwdt 8mo ago|0 comments
12.
CoderForge-Preview: SOTA open dataset for training efficient coding agents(together.ai)
1 points by zagwdt 8mo ago|0 comments
13.
Two years of vector search at Notion: 10x scale, 1/10th cost(notion.com)
2 points by zagwdt 8mo ago|0 comments
14.
Consistency diffusion language models: Up to 14x faster, no quality loss(together.ai)
219 points by zagwdt 8mo ago|96 comments