New Links | ML News

ML Newsnew | past | best | rss | submit

541.	Error Certificates for KV-Cache Eviction via Randomized Design (arxiv.org)
	2 points by sbulaev 70 days ago \| hide \| past \| pdf \| discuss
542.	How Many Bits Can an Adapter Write? Capacity and Memorization of LoRA (arxiv.org)
	2 points by sbulaev 70 days ago \| hide \| past \| pdf \| discuss
543.	Saga: Source Attribution of Generative AI Videos (identifies the model used) (arxiv.org)
	5 points by giuliomagnifico 71 days ago \| hide \| past \| pdf \| discuss
544.	TiRex-2: Generalizing TiRex to Multivariate Data and Streaming (arxiv.org)
	2 points by tosh 71 days ago \| hide \| past \| pdf \| discuss
545.	An Early Warning of Emerging Biosecurity Risks in Frontier LLMs (arxiv.org)
	1 point by StatsAreFun 71 days ago \| hide \| past \| pdf \| discuss
546.	An Exam for Active Observers (arxiv.org)
	2 points by ripbozo 71 days ago \| hide \| past \| pdf \| discuss
547.	The world model remembers, the actor forgets: dissecting AI forgetting on 1 GPU (arxiv.org)
	2 points by gurpnijjer 72 days ago \| hide \| past \| pdf \| discuss
548.	Guardrails as Scapegoats: Auditing Unfaithful Safety Refusals in Tool-Augmented (arxiv.org)
	2 points by sbulaev 72 days ago \| hide \| past \| pdf \| discuss
549.	LLMs can hide text in other text of the same length (arxiv.org)
	5 points by dash2 72 days ago \| hide \| past \| pdf \| discuss
550.	Treating Agent Memory and Costs as Lifecycle and Architecture Problems (arxiv.org)
	2 points by gdad 72 days ago \| hide \| past \| pdf \| discuss
551.	LLMs can write themselves notes to get better at reasoning (arxiv.org)
	1 point by MarcoDewey 72 days ago \| hide \| past \| pdf \| discuss
552.	Emergent Misalignment Recruits a Pre-Existing Persona Subspace (arxiv.org)
	3 points by sbulaev 72 days ago \| hide \| past \| pdf \| discuss
553.	Beyond Static Summarization: Proactive Memory Extraction for LLM Agents (arxiv.org)
	2 points by ankitg12 72 days ago \| hide \| past \| pdf \| discuss
554.	Leaky Language Models: Stealing Architecture and Inference Optimizations (arxiv.org)
	2 points by sbulaev 72 days ago \| hide \| past \| pdf \| discuss
555.	Kat-Coder-v2.5 Technical Report (Qwen3.6-35B-A3B base for post-training) (arxiv.org)
	2 points by truth_seeker 72 days ago \| hide \| past \| pdf \| discuss
556.	Automated sign detection across the Electronic Babylonian Library (arxiv.org)
	6 points by Jimmc414 72 days ago \| hide \| past \| pdf \| discuss
557.	Computational Identifiability (arxiv.org)
	3 points by simonpure 72 days ago \| hide \| past \| pdf \| discuss
558.	Advancing Mathematics Research with AI-Driven Formal Proof Search (arxiv.org)
	2 points by Jimmc414 72 days ago \| hide \| past \| pdf \| discuss
559.	A local RAG assistant for a Romanian government agency, built on an 8GB laptop (arxiv.org)
	2 points by rotariuvladimir 72 days ago \| hide \| past \| pdf \| discuss
560.	Automated Discovery Has No Universally Superior Harness (arxiv.org)
	2 points by matt_d 73 days ago \| hide \| past \| pdf \| discuss
561.	Honey Bee Colony Monitoring via Audio IoT Sensors, Tensorgrams and RNNs (arxiv.org)
	30 points by StatsAreFun 73 days ago \| hide \| past \| pdf \| discuss
562.	Learnability Can Be Independent of ZFC Axioms (2019) (arxiv.org)
	2 points by gone35 74 days ago \| hide \| past \| pdf \| discuss
563.	Backpropagation as Physical Relaxation: Exact Gradients in Finite Time (arxiv.org)
	3 points by binyu 74 days ago \| hide \| past \| pdf \| discuss
564.	SysAdmin: Measuring Instrumental Power-Seeking in Frontier AI (arxiv.org)
	2 points by Brajeshwar 74 days ago \| hide \| past \| pdf \| discuss
565.	Intelligence from Learnable Novelty (arxiv.org)
	6 points by vicgalle_ 74 days ago \| hide \| past \| pdf \| discuss
566.	Measuring Reward-Seeking via Contrastive Belief Updates (arxiv.org)
	1 point by sbulaev 74 days ago \| hide \| past \| pdf \| discuss
567.	A Controlled Study of Attention-Only Transformers (arxiv.org)
	2 points by E-Reverance 74 days ago \| hide \| past \| pdf \| 1 comment
568.	Salience Induction Against Multi-Hop RAG Agents: Threat and Defense (arxiv.org)
	2 points by sbulaev 74 days ago \| hide \| past \| pdf \| discuss
569.	Hardware Mechanisms to Dynamically Throttle AI Performance (arxiv.org)
	2 points by Jimmc414 74 days ago \| hide \| past \| pdf \| discuss
570.	Three-Body Scattering for Generative Modeling (arxiv.org)
	3 points by E-Reverance 74 days ago \| hide \| past \| pdf \| discuss
	More