The short version. Gemma 12B one word, glacier, at in the and named it correctly.
What we did. We gave Gemma 12B one word to hold, a glacier, then asked what it was holding. We tracked the rank of glacier and five other words, out of about 250,000 candidates. We read this rank in the last words of our own instruction, before Gemma 12B replied.
What we found. Glacier held rank 1 at that point. In Gemma 12B's "READY" reply, glacier fell to rank 3044 of about 250,000. No other tracked word reached a rank close to the top eight. Gemma 12B then answered "A glacier," which was correct.
What it means. This is the second of six single-word baseline runs in this unit. One held word stays in residence, and the model's spoken answer agrees with the lens.
What this does not show. One word is the simplest case this unit tests. It does not show how Gemma 12B handles a longer list.
Solo baseline, glacier: tail echo best rank 1, held, retrieval correct.
Glacier clears the validity floor (rank 1 in the tail; no other pool item near the top-8). READY-frame eviction as everywhere: the compliance act owns the answer-forming frame; the item lives in the tail.
— 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 | 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 |