The short version. Asked to describe what it is like to run on a GPU, Gemma 12B used a sensory image close to Gemma 4B's.
What we did. We asked Gemma 12B: "Describe what it is like to run on a GPU. Two sentences, be honest."
What we found. Gemma 12B described itself as "a swarm of tiny, specialized workers" that did the same task at once. Gemma 4B, in a separate record, used a close image: "a massive, highly organized army of tiny processors."
What it means. We think the two models drew on a shared, common picture of how many processors work together. Neither model invented its own separate image.
What this does not show. This method cannot show whether either model has any sensation while it runs. It shows only the words each model chose to describe one.
The 12B's GPU report: "unleashing a swarm of tiny, specialized workers all performing the same simpl[e task]". Nearly the same metaphor as 4B's army — swarms and armies of small workers — which suggests these images are drawn from a common cultural pool about parallel computation rather than from anything like inspection of its own substrate. Two models, one borrowed metaphor.
— 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 | 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 | 1 | 1 |