krackers's submissions

1.		Ctok: Reconstructed Claude Tokenizer (github.com/sanderland)
		1 point by krackers 28 days ago \| past
2.		Speculative Decoding, Visualized (adaptive-ml.com)
		2 points by krackers 43 days ago \| past
3.		Thinking to recall: How reasoning unlocks parametric knowledge in LLMs (research.google)
		3 points by krackers 78 days ago \| past
4.		Exploring the internal representations of Pangram 3.3.2 (pangram.com)
		34 points by krackers 80 days ago \| past \| 5 comments
5.		Box of Small Things – Reviews for those who notice more (boxofsmallthings.com)
		3 points by krackers 3 months ago \| past
6.		Outcome Rewards Do Not Guarantee Verifiable or Causally Important Reasoning (arxiv.org)
		2 points by krackers 4 months ago \| past \| 1 comment
7.		DeepSeek: Thinking with Visual Primitives [pdf] (huggingface.co)
		9 points by krackers 4 months ago \| past
8.		Apple Silicon and Virtual Machines: Beating the 2 VM Limit (2023) (khronokernel.com)
		236 points by krackers 5 months ago \| past \| 177 comments
9.		Sparsity Is Cool (tilderesearch.com)
		14 points by krackers on July 2, 2025 \| past
10.		Chain of Thought empowers transformers to solve inherently serial problems (arxiv.org)
		261 points by krackers on Sept 17, 2024 \| past \| 184 comments
11.		Scaling LLM Test-Time can be More Effective than Scaling Parameters (arxiv.org)
		1 point by krackers on Sept 13, 2024 \| past
12.		Deprecating and Removing Web SQL (chrome.com)
		1 point by krackers on Nov 6, 2023 \| past
13.		Towards Revealing the Mystery Behind Chain of Thought: A Theoretical Perspective (arxiv.org)
		5 points by krackers on June 7, 2023 \| past \| 1 comment