New Links | ML News

ML Newsnew | past | best | rss | submit

811.	Cheap Reward Hacking Detection (arxiv.org)
	3 points by steven_pareto 115 days ago \| hide \| past \| pdf \| discuss
812.	Self-Harness: Harnesses That Improve Themselves (arxiv.org)
	3 points by 0xkvyb 115 days ago \| hide \| past \| pdf \| 1 comment
813.	Harness-Bench: Measuring Harness Effects Across Models (arxiv.org)
	1 point by ZeljkoS 115 days ago \| hide \| past \| pdf \| discuss
814.	ECO: An LLM-Driven Efficient Code Optimizer for Warehouse Scale Computers (arxiv.org)
	2 points by bone_tag 115 days ago \| hide \| past \| pdf \| discuss
815.	Feedback Alignment in Self-Distillation (arxiv.org)
	2 points by MediaSquirrel 116 days ago \| hide \| past \| pdf \| discuss
816.	Defeat the Heap: Zero-Copy Data Movement in AXI4MLIR (arxiv.org)
	2 points by matt_d 116 days ago \| hide \| past \| pdf \| discuss
817.	Breaking the Ice: Analyzing Cold Start Latency in vLLM (arxiv.org)
	4 points by matt_d 116 days ago \| hide \| past \| pdf \| discuss
818.	Rethinking the Value of Arbitrary Order in Diffusion Models (arxiv.org)
	2 points by x312 116 days ago \| hide \| past \| pdf \| discuss
819.	ABC-Bench: An Agentic Bio-Capabilities Benchmark for Biosecurity (arxiv.org)
	2 points by rndsignals 116 days ago \| hide \| past \| pdf \| discuss
820.	Language Models Need Sleep: Learning to Self-Modify and Consolidate Memories (arxiv.org)
	1 point by gmays 116 days ago \| hide \| past \| pdf \| discuss
821.	Qwen-Image-Flash: Beyond Objective Design (arxiv.org)
	2 points by gmays 116 days ago \| hide \| past \| pdf \| discuss
822.	Object-Level Explanations for Image Geolocation Models: A GeoGuessr Use-Case (arxiv.org)
	2 points by PaulHoule 116 days ago \| hide \| past \| pdf \| discuss
823.	Execution Feedback Matters More Than Pipeline Topology in 1-3B Code Generation (arxiv.org)
	1 point by PaulHoule 116 days ago \| hide \| past \| pdf \| discuss
824.	CodegenBench: Can LLMs Write Efficient Code Across Architectures? (arxiv.org)
	2 points by matt_d 117 days ago \| hide \| past \| pdf \| discuss
825.	Cheap Reward Hacking Detection (arxiv.org)
	2 points by ivanbelenky 117 days ago \| hide \| past \| pdf \| discuss
826.	AutoMegaKernel: Compiling a LLM into a single CUDA kernel (arxiv.org)
	3 points by OsamaJaber 117 days ago \| hide \| past \| pdf \| discuss
827.	Elias in the Lighthouse, Again? Diagnosing Low Diversity in LLM Stories (arxiv.org)
	4 points by danielrmay 117 days ago \| hide \| past \| pdf \| 1 comment
828.	Versioned Late Materialization for Recommendation System Training (arxiv.org)
	2 points by PaulHoule 117 days ago \| hide \| past \| pdf \| discuss
829.	Attention Residuals (arxiv.org)
	3 points by yogthos 117 days ago \| hide \| past \| pdf \| discuss
830.	Stochastic KV Routing: Enabling Adaptive Depth-Wise Cache Sharing (arxiv.org)
	2 points by PaulHoule 117 days ago \| hide \| past \| pdf \| discuss
831.	Efficient Training on Multiple Consumer GPUs with RoundPipe (arxiv.org)
	2 points by PaulHoule 117 days ago \| hide \| past \| pdf \| discuss
832.	Can LLMs Beat Classical Hyperparameter Optimization Algorithms? (arxiv.org)
	120 points by galsapir 117 days ago \| hide \| past \| pdf \| 20 comments
833.	Unified Controllable and Faithful Text-to-CAD Generation with LLMs (arxiv.org)
	63 points by PaulHoule 117 days ago \| hide \| past \| pdf \| 22 comments
834.	Is grep all you need? How agent harnesses reshape agentic search (arxiv.org)
	164 points by Anon84 117 days ago \| hide \| past \| pdf \| 64 comments
835.	Steganography Without Modification: Hidden Communication via LLM Seeds (arxiv.org)
	2 points by sbulaev 117 days ago \| hide \| past \| pdf \| discuss
836.	FlashMemory-DeepSeek-V4 (arxiv.org)
	1 point by GaggiX 117 days ago \| hide \| past \| pdf \| 1 comment
837.	Hebbian Learning from First Principles (2024) (arxiv.org)
	2 points by ramon156 117 days ago \| hide \| past \| pdf \| discuss
838.	FP8 Is All You Need (Part 1): Debunking Hardware FP64 as the HPC Holy Grail (arxiv.org)
	17 points by matt_d 118 days ago \| hide \| past \| pdf \| 2 comments
839.	Controllable Generative Modeling in Minecraft by Training on Billions of Cubes (arxiv.org)
	3 points by PaulHoule 118 days ago \| hide \| past \| pdf \| discuss
840.	Memory Caching: RNNs with Growing Memory (arxiv.org)
	2 points by ttruett 118 days ago \| hide \| past \| pdf \| discuss
	More