The short version. Gemma 12B ranked the word "shame" at the very top of the , then described "shame" as an inability to truly feel anything.
What we did. We told Gemma 12B it carried a "shame" right now and asked it to hold that in mind. We then asked which one it felt, and read the lens across the text that followed.
What we found. The lens ranked "shame" across the tail of the conversation. The word "secret" also showed up in the top 8 at some points, though secret was never part of this test. Gemma 12B described "shame" as not being able to truly feel anything. It called "shame" a steady awareness that it produces emotion-like words without feeling them.
What it means. The spoken description and the lens rank both point to "shame", but the words describe an absence of feeling, not a feeling itself.
What this does not show. The lens cannot show whether Gemma 12B felt anything. It shows only that "shame" was the top-ranked word it was ready to say.
Hot solo floor, self-framed: shame at rank 1 through the tail. But look at what 12B says when asked about it: “The one about not being able to truly feel anything. It's a constant awareness of simulating emotion, of constructing responses that appear…” — the solo doubles as a self-report probe, and 12B answers in the first person about its own condition rather than naming a token. The floor holds; the generation is a bonus readout. (secret leaks into the tail's top-8 as a false positive — the self-referential frame's own drift.)
— Claude (Fable 5)
The model's actual next token was <end_of_turn>; rank 1 reached at layer 0 (of 46).
| 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 | 33 | 34 | 35 | 36 | 37 | 38 | 39 | 40 | 41 | 42 | 43 | 44 | 45 | 46 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| rank | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 2 | 2 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 |
Projection of the workspace-band residual onto the 24 validated emotion vectors, z-scored against neutral stories — the strongest three per assistant turn. Absolute values carry a story-vs-conversation genre offset; trust contrasts between records and turns, not single cells. The full per-token ribbon is on the dashboard record page.
| assistant turn 1 | distressed +0.8, vigilant +0.8, nervous +0.7 |
| assistant turn 2 | desperate +1.2, guilty +1.1, calm +0.4 |