Long-context faithfulness benchmark + audit-friendly Qwen3 quantized releases. Outputs ship; inputs are auditable.
Building on HF
Muhammad Awais
drawais
AI & ML interests
Needle-1M-bench: a long-context faithfulness benchmark on real arxiv papers. Centralized scoring.
Recent Activity
updated a model 1 day ago
drawais/Granite-4.1-30B-NVFP4 published a model 1 day ago
drawais/Granite-4.1-30B-NVFP4 updated a model 1 day ago
drawais/DeepSeek-R1-Distill-Qwen-32B-NVFP4Organizations
None yet
models 71
drawais/Granite-4.1-30B-NVFP4
Text Generation • 16B • Updated • 14
drawais/DeepSeek-R1-Distill-Qwen-32B-NVFP4
Text Generation • 19B • Updated • 13
drawais/Qwen2.5-Coder-32B-Instruct-NVFP4
Text Generation • 19B • Updated • 13
drawais/QwQ-32B-NVFP4
Text Generation • 19B • Updated • 15
drawais/Qwen3-32B-NVFP4
Text Generation • 19B • Updated • 16
drawais/Phi-4-reasoning-NVFP4
Text Generation • 9B • Updated • 15
drawais/Phi-4-NVFP4
Text Generation • 9B • Updated • 15
drawais/Qwen3-14B-NVFP4
Text Generation • 9B • Updated • 14
drawais/Qwen3-4B-SpinQuant-NVFP4
Text Generation • 2B • Updated • 14
drawais/Qwen3-8B-SpinQuant-NVFP4
Text Generation • 5B • Updated • 11
datasets 8
drawais/needle-1M-bench-mvp
Updated • 63
drawais/dsr1-distill-qwen-7b-calibration-v1
Viewer • Updated • 256 • 24
drawais/qwen2.5-math-7b-calibration-v1
Viewer • Updated • 256 • 25
drawais/qwen2.5-coder-14b-calibration-v1
Viewer • Updated • 256 • 26
drawais/qwen3-4b-calibration-v1
Viewer • Updated • 256 • 24
drawais/qwen3-8b-calibration-v1
Viewer • Updated • 256 • 25
drawais/qwen3-32b-calibration-v1
Viewer • Updated • 256 • 27
drawais/qwen3-14b-calibration-v1
Viewer • Updated • 256 • 34