Aikido Security Unveils Altar-1 Open-Weight AI for Cybersecurity Defense
2 sources 3 min
AERIAL: Adversarial Evaluation of Robustness in Accuracy-Preserving Low-Precision EEG Decoders
MicroQonv: Reshaping Convolution Tensors for Efficient Microscaling in Training and Inference
abenzerps/Qwen-Image-2.1-GGUF — new model trending #14 on Hugging Face
9 sources 5 min
Train Where the Quantized Model Goes: On-Policy Distillation for Low-Bit Reasoning
Transformers now runs llama.cpp quants
GGUF vs GPTQ vs AWQ vs EXL2: LLM Model Formats Explained (2026)
13 min1
Towards Full Pipeline FP8 Reinforcement Learning for LLMs
PrismML Releases Ternary Bonsai 2 27B: A 5.9 GB Apache 2.0 Model Retaining 98.2% of Qwen3.8 27B Performance
4 min
Shapelearn Qwen 3.8 27B (13.1 GB VRAM)
15 min
PrismML hopes its tiny LLM will change how we all use AI
3 min1
Bonsai 2 27B: Near-Lossless Compression in a 9x Smaller Footprint
Score Centering Stabilizes Off-policy Reinforcement Learning
1
Nunchux AI Introduces VC-Attention: A Training-Free Low-Bit Attention Kernel That Speeds Up Video Diffusion Transformers
5 min1
prism-ml/Ternary-Bonsai-2-27B-gguf — new model trending #29 on Hugging Face
Breaking the 1.58-bit Barrier for Ternary LLMs
Retention-Constrained Post-Training Quantization of Cellpose-SAM for Stem Cell Microscopy
Cactus-Compute/needle3 — new model trending #30 on Hugging Face
6 min
ActionPiece: Rethinking Action Tokenization for Autoregressive Vision-Language-Action Models
1
The Other Half of the Memory Wall: Serving 35B MoEs from SSD with Trained Routing Prediction
Fathom: Per-Query Read Depth for Sparse Decoding over Offloaded KV Caches
1
Privacy-Aligned Personalized Federated Learning with Compact Adaptation and Variable-Length Gaussian Communication
1
VC-Attention: Value Smoothing and Softmax Casting for Low-bit Attention
1
Attention Quantization for Tabular Foundation Models
1