New Links | ML News

ML Newsnew | past | best | rss | submit

1801.	Frontier Models are Capable of In-context Scheming (arxiv.org)
	2 points by william-evans 279 days ago \| hide \| past \| pdf \| 1 comment
1802.	LLM Efficiency: From Hyperscale Optimizations to Universal Deployability (arxiv.org)
	1 point by PaulHoule 279 days ago \| hide \| past \| pdf \| discuss
1803.	The Sparsely-Gated Mixture-of-Experts Layer (2017) [pdf] (arxiv.org)
	1 point by swatson741 279 days ago \| hide \| past \| pdf \| discuss
1804.	Large Language Models Struggle to Learn Long-Tail Knowledge (2023) (arxiv.org)
	1 point by wslh 279 days ago \| hide \| past \| pdf \| discuss
1805.	LLMs, LoRA, and Slerp Shape Representational Geometry of Embeddings (arxiv.org)
	1 point by PaulHoule 280 days ago \| hide \| past \| pdf \| discuss
1806.	Deep sequence models tend to memorize geometrically; it is unclear why (arxiv.org)
	3 points by tzury 280 days ago \| hide \| past \| pdf \| discuss
1807.	Optimal Software Pipelining and Warp Specialization for Tensor Core GPUs (arxiv.org)
	2 points by matt_d 280 days ago \| hide \| past \| pdf \| discuss
1808.	Generative Caching for Structurally Similar Prompts and Responses (arxiv.org)
	1 point by PaulHoule 280 days ago \| hide \| past \| pdf \| discuss
1809.	Propose, Solve, Verify: Self-Play Through Formal Verification (arxiv.org)
	2 points by imakwana 280 days ago \| hide \| past \| pdf \| discuss
1810.	Position: Privacy Is Not Just Memorization (arxiv.org)
	1 point by PaulHoule 280 days ago \| hide \| past \| pdf \| discuss
1811.	A Profit-Based Measure of Lending Discrimination (arxiv.org)
	3 points by neehao 281 days ago \| hide \| past \| pdf \| discuss
1812.	Automating Deception: Scalable Multi-Turn LLM Jailbreaks (arxiv.org)
	3 points by PaulHoule 281 days ago \| hide \| past \| pdf \| discuss
1813.	ChatGPT: Excellent Paper Accept It. Editor: Imposter Found Review Rejected (arxiv.org)
	1 point by belter 281 days ago \| hide \| past \| pdf \| discuss
1814.	Designing Predictable LLM-Verifier Systems for Formal Method Guarantee (arxiv.org)
	59 points by PaulHoule 281 days ago \| hide \| past \| pdf \| 13 comments
1815.	Toward Training Superintelligent Software Agents Through Self-Play SWE-RL (arxiv.org)
	1 point by pama 281 days ago \| hide \| past \| pdf \| discuss
1816.	Towards a Science of Scaling Agent Systems (arxiv.org)
	1 point by Anon84 281 days ago \| hide \| past \| pdf \| discuss
1817.	Beyond Context: Large Language Models Failure to Grasp Users Intent (arxiv.org)
	4 points by mpweiher 281 days ago \| hide \| past \| pdf \| discuss
1818.	Toward Training Superintelligent Software Agents Through Self-Play SWE-RL (arxiv.org)
	2 points by klipt 282 days ago \| hide \| past \| pdf \| discuss
1819.	Prompt Repetition Improves Non-Reasoning LLMs (arxiv.org)
	2 points by ksec 282 days ago \| hide \| past \| pdf \| 1 comment
1820.	Emergent temporal abstractions in autoregressive models enable hierarchical RL (arxiv.org)
	2 points by simonpure 282 days ago \| hide \| past \| pdf \| discuss
1821.	Attention Is Not What You Need: Grassmann Flows as an Attention-Free Alternative (arxiv.org)
	3 points by lexandstuff 283 days ago \| hide \| past \| pdf \| discuss
1822.	Dual Codebook Representationl Learning for Generative Recommendation (arxiv.org)
	2 points by PaulHoule 283 days ago \| hide \| past \| pdf \| discuss
1823.	Yann LeCun: New Vision Language JEPA with Better Performance Than LLMs (arxiv.org)
	10 points by bluedevilzn 283 days ago \| hide \| past \| pdf \| discuss
1824.	Multi-View SVG Generation with Geometric and Color Consistency from a Single SVG (arxiv.org)
	2 points by PaulHoule 283 days ago \| hide \| past \| pdf \| discuss
1825.	Toward Training Superintelligent Software Agents Through Self-Play SWE-RL (Meta) (arxiv.org)
	1 point by xhevahir 283 days ago \| hide \| past \| pdf \| discuss
1826.	LitBench: A Benchmark and Dataset for Reliable Evaluation of Creative Writing (arxiv.org)
	3 points by andy99 284 days ago \| hide \| past \| pdf \| discuss
1827.	Creating General User Models from Computer Use (arxiv.org)
	1 point by handfuloflight 284 days ago \| hide \| past \| pdf \| discuss
1828.	A Scalable Communication Protocol for Networks of Large Language Models (arxiv.org)
	1 point by walterbell 285 days ago \| hide \| past \| pdf \| discuss
1829.	Minimizing Hyperbolic Embedding Distortion with LLM-Guided Hierarchy Structuring (arxiv.org)
	3 points by PaulHoule 286 days ago \| hide \| past \| pdf \| discuss
1830.	Layout-Aware Text Editing for Efficient Conversion of Academic PDFs to Markdown (arxiv.org)
	1 point by 50kIters 286 days ago \| hide \| past \| pdf \| discuss
	More