New Links | ML News

ML Newsnew | past | best | rss | submit

1081.	When innocent tools form dangerous chains to jailbreak LLM agents (arxiv.org)
	2 points by leecoursey 152 days ago \| hide \| past \| pdf \| discuss
1082.	A Theory of Generalization in Deep Learning (arxiv.org)
	4 points by E-Reverance 152 days ago \| hide \| past \| pdf \| discuss
1083.	GLM-5V-Turbo: Toward a Native Foundation Model for Multimodal Agents (arxiv.org)
	163 points by gmays 152 days ago \| hide \| past \| pdf \| 32 comments
1084.	The Last Human-Written Paper: Agent-Native Research Artifacts (arxiv.org)
	2 points by amberjcjj 152 days ago \| hide \| past \| pdf \| discuss
1085.	Faster RL Post-Training Rollouts via System-Integrated Speculative Decoding (arxiv.org)
	1 point by gmays 152 days ago \| hide \| past \| pdf \| discuss
1086.	DeepSeek V4's indexer OOMs at 65K context. We got it to 1M in 6G (arxiv.org)
	8 points by OsamaJaber 152 days ago \| hide \| past \| pdf \| discuss
1087.	Process-Level Reward Modeling for Agentic Data Analysis (arxiv.org)
	4 points by gmays 153 days ago \| hide \| past \| pdf \| discuss
1088.	Transformers Are Inherently Succinct (2025) (arxiv.org)
	62 points by bearseascape 153 days ago \| hide \| past \| pdf \| 9 comments
1089.	Exploring LLM biases to manipulate AI search overview (arxiv.org)
	1 point by Brajeshwar 153 days ago \| hide \| past \| pdf \| discuss
1090.	Hallucination Is Inevitable: An Innate Limitation of Large Language Models (2025) (arxiv.org)
	14 points by drob518 153 days ago \| hide \| past \| pdf \| 11 comments
1091.	Story of Two GPUs: Characterizing the Resilience of Hopper H100 and Ampere A100 (arxiv.org)
	2 points by rbanffy 153 days ago \| hide \| past \| pdf \| discuss
1092.	Iarpa Trojans in Artificial Intelligence (arxiv.org)
	2 points by hlynurd 153 days ago \| hide \| past \| pdf \| discuss
1093.	Learning Randomized Reductions (arxiv.org)
	2 points by matt_d 154 days ago \| hide \| past \| pdf \| discuss
1094.	New research on analyzing and predicting token consumption of coding agents (arxiv.org)
	4 points by jiaxinpei 154 days ago \| hide \| past \| pdf \| 1 comment
1095.	Learning Pseudorandom Numbers with Transformers (arxiv.org)
	11 points by pizza 154 days ago \| hide \| past \| pdf \| 3 comments
1096.	LLMs can hide text in other text of the same length (arxiv.org)
	5 points by m-hodges 155 days ago \| hide \| past \| pdf \| discuss
1097.	A Note on TurboQuant and the Earlier Eden Work (arxiv.org)
	2 points by amitport 155 days ago \| hide \| past \| pdf \| discuss
1098.	Emergent Strategic Reasoning Risks in AI: A Taxonomy-Driven Evaluation Framework (arxiv.org)
	1 point by gmays 155 days ago \| hide \| past \| pdf \| discuss
1099.	AI Self-preferencing in Algorithmic Hiring: Empirical Evidence and Insights (arxiv.org)
	335 points by laurex 155 days ago \| hide \| past \| pdf \| 178 comments
1100.	Refusal in Language Models Is Mediated by a Single Direction (arxiv.org)
	118 points by fagnerbrack 155 days ago \| hide \| past \| pdf \| 45 comments
1101.	Tessera: Unlocking Heterogeneous GPUs Through Kernel-Granularity Disaggregation (arxiv.org)
	1 point by matt_d 156 days ago \| hide \| past \| pdf \| discuss
1102.	Xmemory: Benchmarking Structured AI Memory Against RAG and Hybrid RAG (arxiv.org)
	9 points by alex_petrov 156 days ago \| hide \| past \| pdf \| 2 comments
1103.	Agentic Harness Engineering (arxiv.org)
	15 points by Anon84 157 days ago \| hide \| past \| pdf \| discuss
1104.	MathDuels: Evaluating LLMs as Problem Posers and Solvers (arxiv.org)
	2 points by matt_d 157 days ago \| hide \| past \| pdf \| 1 comment
1105.	Performance Analysis of AI Query Approximation Using Lightweight Proxy Models (arxiv.org)
	1 point by tanelpoder 157 days ago \| hide \| past \| pdf \| discuss
1106.	Kernel Contracts: A Spec. Language for Correctness Across Heterogeneous Silicon (arxiv.org)
	1 point by matt_d 157 days ago \| hide \| past \| pdf \| discuss
1107.	Outcome Rewards Do Not Guarantee Verifiable or Causally Important Reasoning (arxiv.org)
	2 points by krackers 157 days ago \| hide \| past \| pdf \| 1 comment
1108.	Prism: Demystifying Retention and Interaction in Mid-Training (arxiv.org)
	1 point by mdp2021 157 days ago \| hide \| past \| pdf \| discuss
1109.	Fast GPU Linear Algebra via Compile Time Expression Fusion (arxiv.org)
	12 points by matt_d 157 days ago \| hide \| past \| pdf \| discuss
1110.	LLMs Corrupt Your Documents When You Delegate (arxiv.org)
	4 points by uxhacker 157 days ago \| hide \| past \| pdf \| discuss
	More