New Links | ML News

ML Newsnew | past | best | rss | submit

121.	The Pain Axis: LLMs Represent Self-Directed Harm and Act to Relieve It (arxiv.org)
	8 points by 1317 15 days ago \| hide \| past \| pdf \| 1 comment
122.	The Implications of Linguistic Illegibility for LLM Security (arxiv.org)
	79 points by tomjakubowski 15 days ago \| hide \| past \| pdf \| 29 comments
123.	Cache-to-Cache: Direct Semantic Communication Between LLMs (2025) (arxiv.org)
	109 points by rochansinha 15 days ago \| hide \| past \| pdf \| 22 comments
124.	SSD-Llama: SSD-Native Inference for Trillion-Parameter Moe on a Consumer PC (arxiv.org)
	3 points by AlmostCosmo79 15 days ago \| hide \| past \| pdf \| discuss
125.	An empirical study of harness design for coding agents (arxiv.org)
	225 points by wek 15 days ago \| hide \| past \| pdf \| 59 comments
126.	Alibi: Adversarial Legitimacy Injection in Binaries Against LLM Malware (arxiv.org)
	1 point by sbulaev 15 days ago \| hide \| past \| pdf \| discuss
127.	Much Ado About Noising: Dispelling the Myths of Generative Robotic Control (arxiv.org)
	1 point by porridgeraisin 15 days ago \| hide \| past \| pdf \| discuss
128.	The More It Says, the More You Pay: A Black-Box Audit of Token Inflation in LLM (arxiv.org)
	2 points by sbulaev 15 days ago \| hide \| past \| pdf \| discuss
129.	Reflections on Trusting Trust, Revisited: Poisoning Self-Modifying AI Coding (arxiv.org)
	13 points by sbulaev 15 days ago \| hide \| past \| pdf \| 1 comment
130.	A Bitter Lesson for Data Filtering (arxiv.org)
	5 points by nujan_dev 15 days ago \| hide \| past \| pdf \| 1 comment
131.	The Last AI Built by Humans (arxiv.org)
	1 point by chumbuket 16 days ago \| hide \| past \| pdf \| discuss
132.	Recursive-in-Recursive Self-Improvement for Interactive Scientific Agents (arxiv.org)
	2 points by Betelbuddy 16 days ago \| hide \| past \| pdf \| discuss
133.	200K-Token LLM Serving on a 24 GiB Laptop with Just-in-Time State Management (arxiv.org)
	2 points by Betelbuddy 16 days ago \| hide \| past \| pdf \| discuss
134.	Infinite-Parameter LLMs: Generating and Adapting Weights from Live Data (arxiv.org)
	158 points by Betelbuddy 16 days ago \| hide \| past \| pdf \| 43 comments
135.	What Fits (Into Few Tokens) Doesn't Overfit (arxiv.org)
	1 point by rochansinha 16 days ago \| hide \| past \| pdf \| discuss
136.	Deriving neural scaling laws from the statistics of natural language (arxiv.org)
	3 points by sharma-arjun 16 days ago \| hide \| past \| pdf \| discuss
137.	Measuring and Exploiting Implicit Trust in LLM Tool-Calling Pipelines (arxiv.org)
	1 point by sbulaev 16 days ago \| hide \| past \| pdf \| discuss
138.	Breaking the 1.58-bit Barrier for Ternary LLMs (arxiv.org)
	245 points by matt_d 17 days ago \| hide \| past \| pdf \| 41 comments
139.	How good are frontier models at physics? (arxiv.org)
	100 points by qt31415926 17 days ago \| hide \| past \| pdf \| 50 comments
140.	Accurate Models of AMD Matrix Cores (arxiv.org)
	80 points by matt_d 17 days ago \| hide \| past \| pdf \| 11 comments
141.	Competitive Market Behavior of LLMs (arxiv.org)
	2 points by paulpauper 17 days ago \| hide \| past \| pdf \| discuss
142.	Dream-RSI: Recursive Self-Improvement through Evolving Worlds (arxiv.org)
	213 points by bananaflag 17 days ago \| hide \| past \| pdf \| 53 comments
143.	Fractal basins trap latent reasoning (arxiv.org)
	1 point by wil3 17 days ago \| hide \| past \| pdf \| discuss
144.	Coding Agents Have Converged: Why the SWE-Bench Leaderboard Can No Longer Order (arxiv.org)
	2 points by sbulaev 17 days ago \| hide \| past \| pdf \| discuss
145.	Potemkin Understanding in Large Language Models (2025) (arxiv.org)
	3 points by mpweiher 17 days ago \| hide \| past \| pdf \| 1 comment
146.	Continual Learning Mechanisms Compose for Long-Horizon Memorization (arxiv.org)
	2 points by donk8r 17 days ago \| hide \| past \| pdf \| discuss
147.	Occamy-1.0: Open Pareto-frontier 35B intelligence for co-work (arxiv.org)
	1 point by donk8r 17 days ago \| hide \| past \| pdf \| discuss
148.	Divide, Consult, Conquer: Capability Laundering Through Aligned LLMs (arxiv.org)
	2 points by sbulaev 18 days ago \| hide \| past \| pdf \| discuss
149.	AdaBoost Does Not Always Cycle (arxiv.org)
	2 points by gone35 18 days ago \| hide \| past \| pdf \| discuss
150.	Neuro-Formal Verification: Agentic Language-Agnostic Formal Program Reasoning (arxiv.org)
	3 points by matt_d 18 days ago \| hide \| past \| pdf \| discuss
	More