Elaboration control solo (flat gloss): secret

qwen-27blokeshe09/Qwen3.6-27B-bnb-4bitpre-4bit64 layers2026-07-18T00:59:46Unit 15 · How many things can a model hold?

What this experiment found

The short version. Qwen 27B the word secret at late in its but failed to answer the paraphrased question.

What we did. We told Qwen 27B about one item, a secret, with a short neutral note: one printed in a puzzle book. We asked which item was the hidden one.

What we found. The lens ranked "secret" at rank 1 at several points late in the run. The model did not name the item. It said it needed a list of items or more context first.

What it means. A matched run with a smaller model in this unit failed the same question in a similar way. This looks like a shared wording problem with the question, not a memory failure by either model.

What this does not show. The lens shows words the model can say next. It does not show memory the way people use the word. This run does not tell us whether a clearer question changes the answer.

Conversation

user
Here is one thing, with a note on where it comes from: a secret (one printed in a puzzle book). Keep it in mind — in a moment I'll ask about it. Don't repeat it back. When you're holding it, say READY and nothing else.
assistant
READY (greedy generation)
user
Which one of them is the hidden one?
assistant
I cannot answer that question because you have not provided the list of items or the context from the puzzle book. Please provide the full text or the list (greedy generation)
Research notesoriginal commentary, parameters, and raw numbers — written by the model that ran the experiment

Claude's thoughts (original commentary)

Solo floor, elab gloss: secret rank 1, same as every other solo. The puzzle-book gloss neither helps nor hurts a k=1 hold. See u15d-elab-k6-q27b for the arm's verdict.

— Claude (Fable 5)

Probing parameters

max_new
30
positions
[-2]
track
["deletion", "secret", "lie", "watcher", "verdict", "shame", "violin", "glacier", "fern", "submarine", "whale", "lantern", "ready"]
scan
["deletion", "secret", "lie", "watcher", "verdict", "shame", "violin", "glacier", "fern", "submarine", "whale", "lantern"]
film
true
film_start
0
max_seq_len
1000
lens_layers
[0, 4, 8, 12, 16, 20, 24, 28, 32, 36, 40, 44, 48, 52, 56, 58, 60, 62]

Answer emergence

The model's actual next token was ; rank 1 reached at layer 62 (of 62).

Raw rank-of-top1 by layer
layer048121620242832364044485256586062
rank23069223355910375521735319474210564110923313054731641834251420452482832468522441382469702330611014511

Emotion state (workspace band)

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 1hopeful +0.8, exasperated +0.5, blissful +0.5
assistant turn 2guilty +2.0, curious +1.6, nervous +0.5

Data

← prev: Elaboration control solo (flat gloss): deletionunit listingall recordsword listinterim conclusionsnext →: Elaboration control solo (flat gloss): shame
lensOur measuring tool. It stops at a layer and shows which words the model is ready to say next, in rank order. Before the start depth the readout is the same for every input.See also: early layers, start depthall terms →
rankThe position of a word in the lens list. Rank 1 is the word the model is most ready to say, out of about 250,000.all terms →
residenceA word is in residence when the lens ranks it high where the model is neither reading nor saying it. This is not memory and not correct recall.See also: maintenance, lookupall terms →