New Links | ML News

ML Newsnew | past | best | rss | submit

1591.	Code Formatting Silently Consumes Your LLM Budget (arxiv.org)
	1 point by mustaphah 238 days ago \| hide \| past \| pdf \| discuss
1592.	Ernie 5.0 Technical Report (arxiv.org)
	2 points by salkahfi 238 days ago \| hide \| past \| pdf \| discuss
1593.	Causal World Modeling for Robot Control (arxiv.org)
	1 point by mountainview 238 days ago \| hide \| past \| pdf \| discuss
1594.	VL-JEPA: Joint Embedding Predictive Architecture for Vision-Language (arxiv.org)
	2 points by andsoitis 238 days ago \| hide \| past \| pdf \| discuss
1595.	Training Foundation Models Directly on Human Brain Data (arxiv.org)
	1 point by helloplanets 239 days ago \| hide \| past \| pdf \| discuss
1596.	Open Problems in Mechanistic Interpretability (arxiv.org)
	2 points by vinhnx 239 days ago \| hide \| past \| pdf \| discuss
1597.	Psychometric Comparability of LLM-Based Digital Twins (arxiv.org)
	1 point by PaulHoule 239 days ago \| hide \| past \| pdf \| discuss
1598.	AlphaFace: High Fidelity and Real-Time Face Swapper Robust to Facial Pose (arxiv.org)
	1 point by PaulHoule 239 days ago \| hide \| past \| pdf \| discuss
1599.	Horizon-LM: A RAM-Centric Architecture for LLM Training (arxiv.org)
	1 point by chrsw 239 days ago \| hide \| past \| pdf \| discuss
1600.	Show HN: Measuring how AI agent teams improve issue resolution on SWE-Verified (arxiv.org)
	2 points by NBenkovich 239 days ago \| hide \| past \| pdf \| discuss
1601.	Turn-Based Structural Triggers: Prompt-Free Backdoors in Multi-Turn LLMs (arxiv.org)
	2 points by PaulHoule 239 days ago \| hide \| past \| pdf \| discuss
1602.	KV Cache Transform Coding for Compact Storage in LLM Inference (arxiv.org)
	2 points by walterbell 240 days ago \| hide \| past \| pdf \| discuss
1603.	Learning to Reason in 13 Parameters (arxiv.org)
	3 points by nicholascarolan 240 days ago \| hide \| past \| pdf \| discuss
1604.	First Proof (arxiv.org)
	186 points by samasblack 240 days ago \| hide \| past \| pdf \| 122 comments
1605.	Motus: A Unified Latent Action World Model (arxiv.org)
	2 points by mnming 240 days ago \| hide \| past \| pdf \| discuss
1606.	Generative Modeling via Drifting (arxiv.org)
	2 points by fabmilo 240 days ago \| hide \| past \| pdf \| 1 comment
1607.	End-to-End Transformer Acceleration Through Processing-in-Memory Architectures (arxiv.org)
	1 point by PaulHoule 241 days ago \| hide \| past \| pdf \| discuss
1608.	Predicting Zero-Shot Classification Performance for Arbitrary Queries (arxiv.org)
	1 point by PaulHoule 241 days ago \| hide \| past \| pdf \| discuss
1609.	Superposition Yields Robust Neural Scaling (arxiv.org)
	1 point by kioku 241 days ago \| hide \| past \| pdf \| discuss
1610.	First Proof (arxiv.org)
	2 points by amichail 241 days ago \| hide \| past \| pdf \| 1 comment
1611.	LLMs do plan before they genenrate tokens (arxiv.org)
	3 points by kaaaang 241 days ago \| hide \| past \| pdf \| discuss
1612.	AutoOverlap: Enabling Fine-Grained Overlap of Computation and Communication (arxiv.org)
	3 points by matt_d 241 days ago \| hide \| past \| pdf \| discuss
1613.	Axe: A Simple Unified Layout Abstraction for Machine Learning Compilers (arxiv.org)
	1 point by matt_d 241 days ago \| hide \| past \| pdf \| discuss
1614.	Psychometric Jailbreaks Reveal Internal Conflict in Frontier Models (arxiv.org)
	68 points by toomuchtodo 242 days ago \| hide \| past \| pdf \| 60 comments
1615.	SoTA LLM Guardrails by Trusting the Typical [ICLR 2026] (arxiv.org)
	1 point by barthelomew 242 days ago \| hide \| past \| pdf \| discuss
1616.	PeerRank: Autonomous LLM Eval Through Web-Grounded,Bias-Controlled Peer Review (arxiv.org)
	3 points by gaptaclod 242 days ago \| hide \| past \| pdf \| discuss
1617.	The Trigger in the Haystack: Extracting and Reconstructing LLM Backdoor Triggers (arxiv.org)
	1 point by maltalex 242 days ago \| hide \| past \| pdf \| discuss
1618.	Distance Marching for Generative Modeling (arxiv.org)
	4 points by E-Reverance 243 days ago \| hide \| past \| pdf \| discuss
1619.	Recursive Knowledge Synthesis for Multi-LLM Systems (arxiv.org)
	2 points by bob1029 243 days ago \| hide \| past \| pdf \| discuss
1620.	Attention at Constant Cost per Token via Symmetry-Aware Taylor Approximation (arxiv.org)
	164 points by fheinsen 243 days ago \| hide \| past \| pdf \| 96 comments
	More