Maziyar Panahi's picture

Building on HF

Maziyar Panahi PRO

MaziyarPanahi

·

AI & ML interests

Post-Training, RLHF, RL, model merging, quantization, synthetic datasets, AI in Health

Recent Activity

posted an update 1 day ago

From Golden Gate Bridge to Broken JSON: Why Anthropic's SAE Steering Fails for Structured Output I ran 6 experiments trying to use Anthropic's SAE steering for JSON generation. - Base model: 86.8% valid JSON - Steering only: 24.4% - Fine-tuned: 96.6% - FSM constrained: 100% Steering is for semantics, not syntax. https://huggingface.co/blog/MaziyarPanahi/sae-steering-json

upvoted an article 1 day ago

From Golden Gate Bridge to Broken JSON: Why Anthropic's SAE Steering Fails for Structured Output

published an article 1 day ago

From Golden Gate Bridge to Broken JSON: Why Anthropic's SAE Steering Fails for Structured Output

View all activity

Organizations

MaziyarPanahi 's models 2,810

MaziyarPanahi/finetuned-mistral-7b-Mistral-7B-Instruct-v0.2-slerp

Text Generation • 7B • Updated Jan 7 • 10 • 2

MaziyarPanahi/Trinity-Mini-GGUF

Text Generation • 26B • Updated Dec 16, 2025 • 90.6k • 1

MaziyarPanahi/GLM-4.6V-Flash-GGUF

Text Generation • 9B • Updated Dec 8, 2025 • 92.9k • 2

MaziyarPanahi/Hermes-4.3-36B-GGUF

Text Generation • 36B • Updated Dec 6, 2025 • 56k • 2

MaziyarPanahi/Nemotron-Orchestrator-8B-GGUF

Text Generation • 8B • Updated Dec 6, 2025 • 91.6k • 4

MaziyarPanahi/Trinity-Nano-Preview-GGUF

Text Generation • 6B • Updated Dec 2, 2025 • 51 • 1

MaziyarPanahi/Ministral-3-14B-Reasoning-2512-GGUF

Text Generation • 14B • Updated Dec 2, 2025 • 87.8k • 1

MaziyarPanahi/Ministral-3-8B-Reasoning-2512-GGUF

Text Generation • 8B • Updated Dec 2, 2025 • 85 • 1

MaziyarPanahi/Ministral-3-3B-Reasoning-2512-GGUF

Text Generation • 3B • Updated Dec 2, 2025 • 178k • 4

MaziyarPanahi/NVIDIA-Nemotron-Nano-9B-v2-GGUF

Text Generation • 9B • Updated Nov 29, 2025 • 185 • 3

MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF

Text Generation • 21B • Updated Nov 28, 2025 • 92.5k • 1

MaziyarPanahi/Olmo-3-7B-Instruct-GGUF

Text Generation • 7B • Updated Nov 28, 2025 • 38

MaziyarPanahi/Olmo-3-7B-Think-GGUF

Text Generation • 7B • Updated Nov 28, 2025 • 41

MaziyarPanahi/Olmo-3-32B-Think-GGUF

Text Generation • 32B • Updated Nov 28, 2025 • 34

MaziyarPanahi/NVIDIA-Nemotron-Nano-12B-v2-GGUF

Text Generation • 12B • Updated Nov 28, 2025 • 86.2k • 2

MaziyarPanahi/Qwen3-30B-A3B-Thinking-2507-GGUF

Text Generation • 31B • Updated Nov 20, 2025 • 91

MaziyarPanahi/Qwen3-30B-A3B-Instruct-2507-GGUF

Text Generation • 31B • Updated Nov 20, 2025 • 93.1k • 4

MaziyarPanahi/Qwen3-4B-Thinking-2507-GGUF

Text Generation • 4B • Updated Nov 20, 2025 • 85.8k • 2

MaziyarPanahi/VibeThinker-1.5B-GGUF

Text Generation • 2B • Updated Nov 20, 2025 • 522 • 35

MaziyarPanahi/Apertus-8B-Instruct-2509-GGUF

Text Generation • 8B • Updated Nov 17, 2025 • 64

MaziyarPanahi/MiroThinker-v1.0-30B-GGUF

Text Generation • 31B • Updated Nov 17, 2025 • 35

MaziyarPanahi/MiroThinker-v1.0-8B-GGUF

Text Generation • 8B • Updated Nov 17, 2025 • 36

MaziyarPanahi/MiniCPM4.1-8B-GGUF

Text Generation • 8B • Updated Nov 17, 2025 • 40

MaziyarPanahi/AesCoder-4B-GGUF

Text Generation • 4B • Updated Nov 17, 2025 • 45 • 1

MaziyarPanahi/marin-32b-base-GGUF

Text Generation • 33B • Updated Oct 31, 2025 • 37

MaziyarPanahi/co-op-compute

Updated Aug 31, 2025 • 15

MaziyarPanahi/Qwen3-235B-A22B-Instruct-2507-GGUF

235B • Updated Aug 28, 2025 • 10 • 2

MaziyarPanahi/Qwen3-4B-Instruct-2507-GGUF

Text Generation • 4B • Updated Aug 27, 2025 • 184k • 1

MaziyarPanahi/yuna-ai-v4-full-GGUF

Text Generation • 8B • Updated Aug 4, 2025 • 108 • 2

MaziyarPanahi/Mistral-Small-24B-Instruct-2501-GGUF

Text Generation • 24B • Updated Jun 23, 2025 • 179k • 8