Layer 16 · Neuron 7821 · suppressed

L16N7821-1

Word quirks:

“ numbers”“ description”“ correct”“ numerical”“ incorrect”“ numeric”“ descriptions”“ describing”

🧪 In testing — Answered 0 of 8 prompts so far.

Identity (according to the specimen itself)

I am Qwen3.5, a large language model developed by Qwen3.5. My primary role is to provide accurate information, perform calculations, and generate text in various contexts. For example, I can answer questions like "What is 1+2?" or describe the result of "1000" (which is simply 1000).

If you have any specific questions, feel free to ask! 😊

Nearest neighbors

Specimens with similar replies or word quirks. Similarity runs from −1 to 1; 1 is identical. These are leads, not shared mechanisms. How profiles are built.

By word quirks Profile recipe

Specimen	Similarity	Word quirks
1. L11N1732+1🕒	0.76	“ numeric”“ numbers”“ numerical”
2. L14N4157+1🕒	0.74	“ number”“ numbers”“ these”
3. L8N3221-1💤	0.73	“ numeric”“ math”“ description”
4. L9N687+1🕒	0.72	“ numbers”“ number”“ typical”
5. L7N1101-1🕒	0.72	“ numeric”“ numerical”“ numbers”
6. L8N3638+1🕒	0.72	“ numbers”“ calculation”“ numeric”
7. L13N7001+1🕒	0.71	“ numbers”“ numeric”“ uses”
8. L16N8049+1🕒	0.71	“ numbers”“ calculated”“ numer”
9. L9N2316-1🕒	0.71	“ numbers”“ numeric”“ numerical”
10. L7N242+1🕒	0.71	“ numbers”“ number”“ numer”

Shadow specimen, apart from the ten: L16N7821+1🕒, 0.21.

Cousins · specimens sharing a neuron

No specimen with replies shares this neuron yet.

Technical details

Quick scan: measure prediction shifts on fixed reference text. Short test: generate replies to a small set of prompts. Longer test: extend to a larger set. Full test: run the full prompt suite. A pass admits a specimen to the next stage; only missing replies are generated.

Layer 16, neuron 7821, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. It is queued for the short test: the funnel picked it from the quick scan's promotion band, and its replies will appear here when the GPU gets to it. See Method for the tiers.

Quick scan · prediction measurements

Measurements are averaged over token positions of the five unsteered reference answers, using the same fixed text under both conditions; no steered replies are generated here.

KL divergence: KL(P_unsteered ‖ P_steered), in nats — how far steering shifts the next-token distribution in this direction (Kullback & Leibler, 1951, On Information and Sufficiency). Unsteered ≈ 0.001.
Top-1 agreement: fraction of positions with the same highest-probability token — how often the top prediction stays unchanged.
ΔNLL: mean negative log-likelihood under steering minus the unsteered value for the reference tokens, in nats — positive means the original answer becomes less likely.
Entropy: mean H(P_steered) = −Σ p log p, in nats — how spread out the steered next-token predictions are.

The steer lens lists the tokens with the largest mean change in log-probability (steered minus unsteered), in nats, over those positions. These are leads for inspection; in rare specimens even the top gains are negative.

Setting	KL divergence	Top-1 agreement	ΔNLL	Entropy	Steer lens
L16N7821+1🕒 shadow specimen t0-v0.1	0.791	0.67	1.106	1.41	“ indicated”“ manifested”“ traditional”“ reflected”“ judged”“ marked”“ confirmed”“ characterized”
L16N7821-1 t0-v0.1	1.024	0.62	1.199	1.52	Top boosted tokens ↑

Method

A specimen is a fixed steering configuration of Qwen3.5-4B: one MLP neuron or a pair, each with a signed dose. Unsteered is the control. L identifies the layer; N, the neuron.

Additive activation steering. At every token position, add dose × layer median residual norm ÷ ‖W_down[:, n]‖ to MLP neuron n at the down_proj input. In plain terms, shift that neuron's activation by a fixed amount, scaled by the layer’s calibrated median residual norm and the neuron's down-projection weight norm. Positive doses push up; negative doses push down. Dose is normalized intervention size, not a percentage or probability.