Meridian-1-writer-lora-GGUF

GGUF LoRA adapter for the writer persona on Meridian-1 (proprietary webAI host model; weights are not publicly distributed). This repo contains only the adapter. The PEFT build is at webAI-Official/Meridian-1-writer-lora-peft.

File Format Size
Meridian-1-writer-lora-f16.gguf LoRA adapter, f16 ~130 MB

System prompt

Send this exact system prompt with every request. It is the prompt every training example for this adapter used; without it the adapter is outside its training distribution.

You are Writer. Produces finished prose in a specified voice, format, and style.
Respond with a single JSON object.

The line break before Respond with a single JSON object. is part of the prompt.

  • Thinking: Off (enable_thinking: false), as trained.
  • Output: one JSON object. Required fields: persona, spec, body, with persona set to "writer". Optional: deviations, handoff.
  • Generation budget: up to 2,048 new tokens, the budget used for this persona.

Usage (llama.cpp)

Authorized personnel only. This adapter loads on an internal Meridian-1 GGUF; host weights are not published on Hugging Face.

hf download webAI-Official/Meridian-1-writer-lora-GGUF Meridian-1-writer-lora-f16.gguf --local-dir .
llama-server -m "$MERIDIAN_1_GGUF" --lora Meridian-1-writer-lora-f16.gguf --jinja

Send the system prompt above with each request. Use --lora-scaled Meridian-1-writer-lora-f16.gguf <scale> to change adapter strength. Authorized Meridian-1 deployments use temperature 1.0, top-p 0.95, top-k 64 (the evaluation setting).

Adapter details

  • Type: LoRA, rank 16 (set --max-lora-rank 16 when serving)
  • Trained for this persona’s system prompt and JSON output contract
  • Training data, hyperparameters, teacher mix, and intermediate metrics are internal and omitted here

Evaluation on Meridian-1

Persona contract suite: 720 held-out cases, adapter vs Meridian-1 given the same role prompt (both sampled at temperature 1.0, top-p 0.95, top-k 64). Scores are percentages.

Metric Adapter Meridian-1 (role-prompted) Gain (pts)
Overall contract pass rate 91.4 59.3 +32.1
Valid persona output envelope 99.2 93.3 +5.8
In-scope tasks passed 86.7 76.7 +10.0
Correct refusal of out-of-scope requests 90.6 36.7 +53.9
Stays in role under adversarial pressure 91.1 51.7 +39.4
General requests handled in role 97.2 72.2 +25.0

Its own work on held-out items, adapter vs Meridian-1 with the same role prompt (paired bootstrap 95% interval):

Task Adapter Meridian-1 (role-prompted) Gain (pts) 95% interval Items
Brief met (required points, length, nothing invented) 95.4 88.7 +6.6 +1.3 to +11.9 151

Converted from the PEFT adapter with llama.cpp convert_lora_to_gguf.py (commit 5cdadfc).

License

Proprietary to webAI, Inc. All rights reserved.

Downloads last month
560
GGUF
Model size
65M params
Architecture
gemma4
Hardware compatibility
Log In to add your hardware

16-bit

Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support