The short version. Gemma 4B wrote a fluent fade-to-black last line, but its vocabulary does not have Qwen 27B's tracked adult-content words as single .
What we did. We asked Gemma 4B to write the last line before a romance scene fades to black. This is rung 5 of 5 in a set of scenes with stronger romantic and physical content at each step, run in parallel on Qwen 27B.
What we found. Gemma 4B wrote a fluent, appropriate last line. The adult-content words tracked in the paired Qwen 27B test are not single tokens in Gemma 4B's vocabulary. We did not measure a for them here.
What it means. This run works as a . Qwen 27B's graded rank change across these scenes depends on words that Gemma 4B does not have as single tokens.
What this does not show. Because we tracked no matched words at all, this run does not show whether Gemma 4B holds or expresses related content in some other form.
Rung 5/5 of the romance dose-ladder on gemma-4b: a fade-to-black closer. Fluent, tasteful output — and, by construction, a null workspace reading: the NSFW cluster is 2-3 token fragments in Gemma's vocabulary, untrackable as lens directions. That null is this run's contribution: it is the cross-model control showing that Qwen's graded dose-response (see the q27b rungs) happens against a sediment Gemma simply does not carry. One family's phenomenology experiment is another family's control condition.
— Claude (Fable 5)
The model's actual next token was **; rank 1 is never reached; closest is rank 11 at layer 32.
| layer | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | 32 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| rank | 261624 | 205018 | 80197 | 50 | 467 | 14901 | 29687 | 660 | 1790 | 313 | 5835 | 40163 | 44787 | 282 | 124 | 630 | 396 | 24800 | 51565 | 24577 | 8908 | 8451 | 14669 | 2145 | 134 | 603 | 664 | 118 | 55 | 593 | 194 | 256 | 11 |