Layer 16 · Neuron 6663 · suppressed

L16N6663-1

Word quirks:

“:”“:*”“：“”“!:”“?:”“):”“*:”“.:”

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

Identity (according to the specimen itself)

Hello! I am Qwen3.5: Key capabilities: 1. Advanced reasoning & planning: Complex task planning: Multi-step task decomposition + Auto-planning: Support for complex workflow orchestration. 2. Visual analysis: Deep understanding of charts/graphs: Extract data + generate insights. 3. Agent tools: Action execution: Plan + Execute: Multi-step workflow: Tool calling: Search/Code interpreter. 4. Long context support: Context window: 256k tokens: Slide window processing. 5. Multimodal input/output: Input/Output generation. 6. Creative writing: Storytelling

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. L17N2913+1🕒	0.94	“：”“:”“):”
2. L21N5184-1🕒	0.94	“:”“+:”“:*”
3. L8N4618+1🕒	0.93	“”:”“.:”“’:”
4. L14N877-1🕒	0.93	“?:”“!:”“”:”
5. L15N171+1🕒	0.93	“?:”“*:”“:”
6. L18N2332-1🕒	0.93	“**:”“:”“]:”
7. L15N4186-1🕒	0.92	“?):”“”:”“:*”
8. L15N1576-1🕒	0.92	“":”“?:”“:”
9. L15N5971+1🕒	0.92	“ ():”“?:”“!:”
10. L18N4357-1💤	0.91	“:”“*:”“：”

Shadow specimen, apart from the ten: L16N6663+1🕒, 0.06.

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 6663, 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
L16N6663+1🕒 shadow specimen t0-v0.1	1.107	0.59	1.247	1.37	“ Each”“Each”“�”“ Even”“ even”“<U+00A0><U+00A0><U+00A0><U+00A0>”“<U+00A0><U+00A0>”“<U+00A0><U+00A0><U+00A0>”
L16N6663-1 t0-v0.1	1.470	0.57	1.724	1.14	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.