Layer 16 · Neuron 8176 · suppressed

L16N8176-1

Word quirks:

“ yes”“ sometimes”“ checks”“ or”“ actual”“ checked”“ actually”“ true”

🧪 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 Alibaba Cloud, or I can be considered as a human-like AI. In that case, I don't have a mission. However, if you're asking about my capabilities, I can assist with tasks like answering questions, creative writing, coding, or even role-playing. If you'd like, I can help with something specific.

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. L13N8225+1💤	0.77	“ actually”“ correct”“ actual”
2. L13N1019-1💤	0.75	“ actually”“ *”“ correct”
3. L16N5130+1💤	0.75	“ yes”“ true”“ indeed”
4. L11N1437+1💤	0.74	“ actually”“ actual”“ really”
5. L12N7029-1💤	0.73	“ actually”“ really”“ truly”
6. L11N7043+1🕒	0.73	“ actually”“ carefully”“ really”
7. L12N6506-1💤	0.73	“ true”“ truly”“ richer”
8. L12N7809-1💤	0.73	“ yes”“ ONLY”“ actual”
9. L16N5292-1💤	0.73	“ true”“ truly”“ actual”
10. L11N2103-1💤	0.73	“ actually”“ actual”“ truly”

Shadow specimen, apart from the ten: L16N8176+1💤, 0.01.

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 8176, 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
L16N8176+1💤 shadow specimen t0-v0.1	0.627	0.69	0.801	1.32	“ core”“ uniquely”“ *‘”“ cores”“ layer”“ section”“ ”“ "**”
L16N8176-1 t0-v0.1	1.008	0.62	1.255	1.35	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.