Meridian-1-writer-lora-GGUF

Proprietary and Internal to webAI, Inc. — Confidential. All rights reserved. No license granted; authorized personnel only.

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 and Internal to webAI, Inc. — Confidential. All rights reserved. No license granted; authorized personnel only.

Meridian-1 is proprietary to webAI, Inc. No license to Meridian-1 is granted by this adapter card or by publication of this adapter.