New Links | ML News

ML Newsnew | past | best | rss | submit

271.	Reward Guided Speculative Decoding (arxiv.org)
	1 point by E-Reverance 34 days ago \| hide \| past \| pdf \| discuss
272.	Evaluating Skills, Not Just Agents: Agentic Continuous Evaluation of Skills (arxiv.org)
	5 points by swolpers 34 days ago \| hide \| past \| pdf \| discuss
273.	Scaling Domain Data Repetition in LLM Pretraining (arxiv.org)
	1 point by gmays 34 days ago \| hide \| past \| pdf \| discuss
274.	DumpsterCluster: From Dumpster Diving to Serving Llama-70B on $60 GPUs (arxiv.org)
	2 points by bookofjoe 34 days ago \| hide \| past \| pdf \| 1 comment
275.	FreeToken: Efficient Edge-Native Moe Serving with Bandwidth-Adaptive Execution (arxiv.org)
	2 points by gmays 34 days ago \| hide \| past \| pdf \| discuss
276.	Deep Learning Is Not So Mysterious or Different (arxiv.org)
	2 points by Anon84 34 days ago \| hide \| past \| pdf \| discuss
277.	PLCBench: Can Autonomous LLM Agents Turn PLC Access into Sustained Physical (arxiv.org)
	2 points by sbulaev 35 days ago \| hide \| past \| pdf \| discuss
278.	Full-Bandwidth Transformer (arxiv.org)
	1 point by gmays 35 days ago \| hide \| past \| pdf \| discuss
279.	Eight Things to Know about Large Language Models(2023) (arxiv.org)
	2 points by turingbook 35 days ago \| hide \| past \| pdf \| 1 comment
280.	SwarmWorld: Stigmergic technological evolution in societies of LLM agents (arxiv.org)
	3 points by wintercarver 35 days ago \| hide \| past \| pdf \| 1 comment
281.	30% on AR-AGI-1 at $0.0007 per task (arxiv.org)
	3 points by randomizedalgs 35 days ago \| hide \| past \| pdf \| discuss
282.	WikiSkill: Compiling Agent Experience for Skill Evolution (arxiv.org)
	1 point by joshcorbin 35 days ago \| hide \| past \| pdf \| discuss
283.	Beyond the Editing Canvas: Evidence Divergence in Ooxml-to-LLM Ingestion (arxiv.org)
	3 points by sbulaev 36 days ago \| hide \| past \| pdf \| discuss
284.	Safety Does Not Compose: Non-Decaying Loop State for Autonomous LLM Agents (arxiv.org)
	3 points by sbulaev 36 days ago \| hide \| past \| pdf \| discuss
285.	Why not to use the Gaussian kernel (arxiv.org)
	2 points by E-Reverance 36 days ago \| hide \| past \| pdf \| discuss
286.	Autonomous Mathematical Discovery in an Open-World Multi-Agent Environment (arxiv.org)
	122 points by stephenchung 36 days ago \| hide \| past \| pdf \| 40 comments
287.	Compiling Agent Experience into Persistent Knowledge for Skill Evolution (arxiv.org)
	7 points by tcp_handshaker 36 days ago \| hide \| past \| pdf \| discuss
288.	LLMs Can Design Near-Optimal OR Algorithms (arxiv.org)
	4 points by tcp_handshaker 36 days ago \| hide \| past \| pdf \| discuss
289.	AutoSaddler: Automatic Harness Optimization (arxiv.org)
	23 points by drseu55 36 days ago \| hide \| past \| pdf \| 1 comment
290.	Difference-in-Differences on a Censored Rating Scale Can Manufacture an Effect (arxiv.org)
	2 points by sbulaev 36 days ago \| hide \| past \| pdf \| discuss
291.	Anchoring Bias in LLM-as-a-Judge Systems: Prior Scores Compromise Evaluation (arxiv.org)
	1 point by sbulaev 37 days ago \| hide \| past \| pdf \| discuss
292.	GenRec: An LLM-Backed Recommendation Ranker at Netflix (arxiv.org)
	2 points by armcat 37 days ago \| hide \| past \| pdf \| discuss
293.	WAFT: Warping-Alone Field Transforms for Optical Flow (2025) (arxiv.org)
	2 points by peter_d_sherman 37 days ago \| hide \| past \| pdf \| 1 comment
294.	Evomal: Self-Poisoning in Self-Evolving Coding Agents (arxiv.org)
	1 point by sbulaev 37 days ago \| hide \| past \| pdf \| discuss
295.	Flare: Verifying MILP Reformulations with LLM-Based Theorem Proving (arxiv.org)
	2 points by henryrobbins00 37 days ago \| hide \| past \| pdf \| discuss
296.	Foundation Model 3x better at predicting cancer treatment (arxiv.org)
	3 points by shcheklein 37 days ago \| hide \| past \| pdf \| discuss
297.	Flint: Efficiently Leveraging High Bandwidth Flash for LLM Inference (arxiv.org)
	2 points by metrofun 37 days ago \| hide \| past \| pdf \| discuss
298.	Noise-Driven Escape from Metastable Phases Explains Grokking in DNNs (arxiv.org)
	2 points by jerlendds 38 days ago \| hide \| past \| pdf \| 4 comments
299.	Metaⁿ: Recursive Self-Improvement Through Emergent Depth (arxiv.org)
	1 point by sbulaev 38 days ago \| hide \| past \| pdf \| discuss
300.	Prefix Sliding for efficient test-time scaling (arxiv.org)
	2 points by E-Reverance 38 days ago \| hide \| past \| pdf \| discuss
	More