Все публикации

0:18:00

xLAM: A Family of Large Action Models to Empower AI Agent Systems

0:07:33

In Defense of RAG in the Era of Long-Context Language Models

0:08:56

[QA] Building Math Agents with Multi-Turn Iterative Preference Learning

0:35:53

Building Math Agents with Multi-Turn Iterative Preference Learning

0:26:54

Attention Heads of Large Language Models: A Survey

0:08:12

[QA] Attention Heads of Large Language Models: A Survey

0:07:58

[QA] The AdEMAMix Optimizer: Better, Faster, Older

0:17:35

The AdEMAMix Optimizer: Better, Faster, Older

0:08:28

[QA] Planning In Natural Language Improves LLM Search For Code Generation

0:26:10

Planning In Natural Language Improves LLM Search For Code Generation

0:13:48

MMMU-Pro: A More Robust Multi-discipline Multimodal Understanding Benchmark

0:10:58

[QA] MMMU-Pro: A More Robust Multi-discipline Multimodal Understanding Benchmark

0:12:00

Sample what you can't compress

0:07:58

[QA] Sample what you can't compress

0:09:00

[QA] CONTEXTCITE: Attributing Model Generation to Context

0:07:18

CONTEXTCITE: Attributing Model Generation to Context

0:15:41

FLUX that Plays Music

0:07:45

[QA] FLUX that Plays Music

0:06:52

Modularity in Transformers: Investigating Neuron Separability & Specialization

0:18:20

Smaller, Weaker, Yet Better: Training LLM Reasoners via Compute-Optimal Sampling

0:07:47

[QA] Smaller, Weaker, Yet Better: Training LLM Reasoners via Compute-Optimal Sampling

0:17:16

Dolphin: Long Context as a New Modality for Energy-Efficient On-Device Language Models

0:08:10

[QA] Dolphin: Long Context as a New Modality for Energy-Efficient On-Device Language Models

0:09:25

CycleGAN with Better Cycles