zunhaisu

zunhai's activity

• All
• Models
• Datasets
• Spaces
• Buckets
• Papers
• Collections
• Community
• Posts
• Upvotes
• Likes
• Articles

upvoted a collection 5 days ago

StartLux-Decision

Collection

Typed decision models from 0.8B to 27B. Every question comes back with a probability for each option. • 21 items • Updated 4 days ago • 7

upvoted a collection about 1 month ago

Massive Activations in Hybrid Linear Attention Models

Collection

Official PAS, ISP, gating-ablation, and scale-study checkpoints for Massive Activations in Hybrid Linear Attention Models. • 10 items • Updated Aug 13 • 2

upvoted a paper about 2 months ago

LEGO-RL: Harness-Native Reinforcement Learning for Coding Agents

Paper • 2608.17393 • Published Aug 18 • 26

upvoted a collection about 2 months ago

Lego-RL

Collection

Harness-native RL for coding agents: the trained policy and the training task index. • 9 items • Updated 21 days ago • 11

upvoted 2 papers about 2 months ago

Massive Activations in Hybrid Linear Attention Large Language Models: Pre-Attention Spikes and Inter-Spike Plateaus

Paper • 2608.12149 • Published Aug 12 • 30

UniMoMo: Expert Merging-Based MoE Acceleration for Large Recommendation Models

Paper • 2608.08627 • Published Aug 9 • 13

upvoted a collection 5 months ago

WTF GENIUS PAPERS

Collection

Papers that made me appreciate my major and my life a little more. obs=Observation, innov=Innovation. Most papers are abt improving tiny models. • 422 items • Updated 1 day ago • 95

upvoted a paper 5 months ago

OScaR: The Occam's Razor for Extreme KV Cache Quantization in LLMs and Beyond

Paper • 2605.19660 • Published May 19 • 39

upvoted a paper 6 months ago

LARY: A Latent Action Representation Yielding Benchmark for Generalizable Vision-to-Action Alignment

Paper • 2604.11689 • Published Apr 13 • 21

upvoted a collection 6 months ago

cool-papers

Collection

122 items • Updated 17 days ago • 11

upvoted 2 papers 6 months ago

Attention Sink in Transformers: A Survey on Utilization, Interpretation, and Mitigation

Paper • 2604.10098 • Published Apr 11 • 84

SpecEyes: Accelerating Agentic Multimodal LLMs via Speculative Perception and Planning

Paper • 2603.23483 • Published Mar 24 • 62

upvoted a paper 8 months ago

Scaling Embeddings Outperforms Scaling Experts in Language Models

Paper • 2601.21204 • Published Jan 29 • 107

upvoted 2 papers 9 months ago

Locate, Steer, and Improve: A Practical Survey of Actionable Mechanistic Interpretability in Large Language Models

Paper • 2601.14004 • Published Jan 20 • 49

SWE-Lego: Pushing the Limits of Supervised Fine-tuning for Software Issue Resolving

Paper • 2601.01426 • Published Jan 4 • 25

upvoted 2 papers 11 months ago

Unveiling Super Experts in Mixture-of-Experts Large Language Models

Paper • 2507.23279 • Published Jul 31, 2025 • 1

DoPE: Denoising Rotary Position Embedding

Paper • 2511.09146 • Published Nov 12, 2025 • 76

upvoted a paper about 1 year ago

Gated Attention for Large Language Models: Non-linearity, Sparsity, and Attention-Sink-Free

Paper • 2505.06708 • Published May 10, 2025 • 13