Skip to content

jeeves:9b

2 TagsUpdated 9B params8192 contextEnglishApache-2.0by PostHog

Jeeves-9B without thinking, with the fitted temperature 1.86. Needs a 24 GB GPU.

9b
ollaya run jeeves:9b --preset triage "I was charged twice for my subscription this month and want a refund."

Details

  • graphd17189240420 · 11 MBonnx · jeeves · 9B · fp32
  • weightsc8d38178b51c · 4.3 GBhuggingface.co/PostHog/jeeves/resolve/8622b7d…/model-00001-of-00005.safetensors
  • weightsa25e1a608a5b · 4.3 GBhuggingface.co/PostHog/jeeves/resolve/8622b7d…/model-00002-of-00005.safetensors
  • weightsab693e0b5940 · 4.2 GBhuggingface.co/PostHog/jeeves/resolve/8622b7d…/model-00003-of-00005.safetensors
  • weights9942b7369222 · 3.1 GBhuggingface.co/PostHog/jeeves/resolve/8622b7d…/model-00004-of-00005.safetensors
  • weightsdd63614f1dc8 · 2.0 GBhuggingface.co/PostHog/jeeves/resolve/8622b7d…/model-00005-of-00005.safetensors
  • weights2ab935dea9ee · 8 MBhuggingface.co/PostHog/jeeves/resolve/8622b7d…/head.pt
  • tokenizer06b9509352d2 · 20 MBhuggingface.co/jaredpalmer/kev-9b/resolve/2629c06…/tokenizer.json
  • decision96633b0ae45b · 2 KB{"engine": "onnx", "family": "jeeves", "encoder": "", "layout": "jeeves-markers-v1", …}
  • calibrationb3c01d4ac3e3 · 217 B{"temperature": [1.859, 1.859, 1.859]}
  • license257d23f4db37 · 10 KBJeeves-9B by PostHog (https://huggingface.co/PostHog/jeeves, https://github.com/PostHog/jeeves): Qwen3.5-9B with its LoRA merged, and a pointer head, Apache-2.0.

Every layer is checked against its sha256 when it is pulled. Weights and tokenizers download from the model author's Hugging Face repository at a pinned commit; Ollaya never re-hosts them.

Jeeves readme and all models