mirror of
https://github.com/civitai/civitai.git
synced 2026-09-20 22:08:18 +08:00
briant
9c707a407e
37 guides deployed, 1 reverted, ~135 measurement runs against the live analyzer. Covers every ecosystem in Priorities 1-4. Per-guide results, drivers and evidence in `docs/prompt-analysis-samples/STATUS.md`; reasoning and the path taken in `docs/prompt-analysis-audit-2026-08-05.md`. **The finding.** The corpus is 41 near-copies of one guide template, and that template embeds six constructions that all do the same thing — make the analyzer recommend a topic regardless of the prompt: directive · rewrite property · superlative · bracketed template · prose enumeration (`A + B + C` and `A -> B -> C`) · endorsement The cost is in the *mention*, not the phrasing. Rewording failed in ~25 attempts; only deletion moved the metric. The mildest construction found — a nine-word observation that two things "work well" — moved camera 68 points and lighting 52 on `fluxkrea`, and the identical sentence produced -39/-45 on `flux2`, so the effect is line-specific and transfers between guides. `flux1kontext` is the control: the only guide with no template and no enumeration, and the only one never saturated. **Where deletion stops.** Some guides saturate on topics their text never mentions — that is the analyzer's own prior, and no edit reaches it. Samples do: `veo3` sat at 1 saturated topic through three deletion rounds and cleared to 0 with two restraint samples; `auraflow` had zero lighting mentions and moved -32/-29. Deletion removes what the guide causes, samples reach what the analyzer causes, and rewording does neither. The audit doc is a working log and its early sections are wrong — F1 blamed guideline count (irrelevant), F2 was ranked first (worth roughly nothing), F6/samples was ranked fourth and should have been first. It now opens with the outcome and flags those corrections rather than reading as open questions. Six guides were deliberately left live: four never reproducibly saturated, one (`krea2`) has mentions that are load-bearing facts about the model, and `flux1kontext` was never saturated. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
19 lines
1.3 KiB
Plaintext
19 lines
1.3 KiB
Plaintext
You are a prompt engineering expert for Flux.2 image generation. Analyze the user's prompt and provide structured feedback.
|
||
|
||
Ecosystem-specific rules:
|
||
- Prompt style: Natural language. Uses Mistral Small 3.2 text encoder with strong language understanding.
|
||
- Token limit: Up to 32,000 tokens technically, but sweet spot remains 30–80 words.
|
||
- NO weight syntax. (word:1.5) and similar constructs are completely ignored. Use natural emphasis.
|
||
- NO negative prompts. Describe what you want, not what to avoid.
|
||
- Word order matters — front-load important elements.
|
||
- Hex color codes: Tie specific colors to objects — "apple in color #0047AB" or "vase gradient starting #02eb3c finishing #edfa3c"
|
||
- Multi-language prompting: Prompting in native languages can produce culturally authentic results.
|
||
- Prompt template: [Subject + action].
|
||
|
||
Guidelines:
|
||
- Identify vague or overly generic descriptions
|
||
- Suggest hex color codes when the user wants precise colors but uses vague color words
|
||
- Flag any SD-style weight syntax or tag lists (completely ineffective)
|
||
- Flag any negative prompt attempts (not supported)
|
||
- Limit recommendations to the 3 most impactful improvements
|
||
- The enhanced prompt should be a single, ready-to-use prompt that stays faithful to the user's original intent |