MNeMoNiCuZandClaude Opus 5.5
92a58d6774
- Migrate all nodes to the V3 node API; old workflows migrate automatically - Add help pages for every node, opened from a ? icon in the title bar - Minor fixes and tooltip, placeholder and default improvements - Bump version to 3.0.0 Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
1.1 KiB
1.1 KiB
⛔ Generate Negative Prompt
Experimental. Generates a negative prompt from a positive one, using a small GPT-2 model fine-tuned on Civitai negative prompts.
The model is weak and largely random: treat the output as a starting point, not a recommendation. NSFW terms can appear.
Inputs
- input_prompt — The positive prompt to react to.
- max_length — Token cap on the generated text.
- num_beams — Beam search width. Higher is slower and more conservative.
- temperature — Randomness. Lower is more predictable.
- top_k — Consider only the K most likely next words. Lower is more predictable, higher is more varied.
- top_p — Nucleus sampling cutoff. Lower keeps output focused.
- blocked_words — One entry per line; each is removed from the result. Handy for stripping embeddings you do not have installed.
Outputs
- negative_prompt — The generated text.
Notes
The model files live in nodes/negativeprompt/ and are loaded on every run, so
the first call after a fresh start is slow. The node is not an output node, so
it only runs when something downstream needs its result.