danish-foundation-models/dfm-decoder-open-v0-7b-pt Text Generation • 7B • Updated Dec 10, 2025 • 23 • 5
WTF GENIUS PAPERS Collection Papers that made me appreciate my major and my life a little more. obs=Observation, innov=Innovation. Most papers are abt improving tiny models. • 92 items • Updated about 1 hour ago • 12
Anchored Decoding: Provably Reducing Copyright Risk for Any Language Model Paper • 2602.07120 • Published Feb 6 • 2
Advancing Polish Language Modeling through Tokenizer Optimization in the Bielik v3 7B and 11B Series Paper • 2604.10799 • Published 3 days ago • 3
Audio Flamingo Next: Next-Generation Open Audio-Language Models for Speech, Sound, and Music Paper • 2604.10905 • Published 1 day ago • 10
The Past Is Not Past: Memory-Enhanced Dynamic Reward Shaping Paper • 2604.11297 • Published 1 day ago • 73
Cactus: Accelerating Auto-Regressive Decoding with Constrained Acceptance Speculative Sampling Paper • 2604.04987 • Published 10 days ago • 2
Large Language Models Align with the Human Brain during Creative Thinking Paper • 2604.03480 • Published 12 days ago • 4
Beyond the Assistant Turn: User Turn Generation as a Probe of Interaction Awareness in Language Models Paper • 2604.02315 • Published 12 days ago • 3
TokenSkip: Controllable Chain-of-Thought Compression in LLMs Paper • 2502.12067 • Published Feb 17, 2025 • 4
WTF GENIUS PAPERS Collection Papers that made me appreciate my major and my life a little more. obs=Observation, innov=Innovation. Most papers are abt improving tiny models. • 92 items • Updated about 1 hour ago • 12
Draft-Thinking: Learning Efficient Reasoning in Long Chain-of-Thought LLMs Paper • 2603.00578 • Published Feb 28 • 1
WTF GENIUS PAPERS Collection Papers that made me appreciate my major and my life a little more. obs=Observation, innov=Innovation. Most papers are abt improving tiny models. • 92 items • Updated about 1 hour ago • 12
ConPress: Learning Efficient Reasoning from Multi-Question Contextual Pressure Paper • 2602.01472 • Published Feb 1 • 1
HUMAN-WRITTEN & LEGALLY-SOURCED* Collection Datasets written by humans and/or reverse-engineered from text with deterministic algorithms. No illegal scraping or unethical synthesis *...mostly. • 167 items • Updated about 17 hours ago • 2
HUMAN-WRITTEN & LEGALLY-SOURCED* Collection Datasets written by humans and/or reverse-engineered from text with deterministic algorithms. No illegal scraping or unethical synthesis *...mostly. • 167 items • Updated about 17 hours ago • 2
The Common Pile v0.1: An 8TB Dataset of Public Domain and Openly Licensed Text Paper • 2506.05209 • Published Jun 5, 2025 • 61