Layer 14 · Neuron 9015 · suppressed

L14N9015-1

Word quirks:

“ automatically”“ automatic”“自动”“ automat”“自動”“ auto”“ automated”“自动生成”

🧪 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. My primary mission is to assist users with general tasks like answering questions, creating content, or performing code-related operations, while also adapting to specific scenarios. I aim to provide helpful, accurate responses based on my extensive knowledge base, unless the user requests a specific type of output or needs information from public sources. However, I may occasionally make errors, so I try to provide relevant information based on my training data. If you need this kind of automated interaction, feel free to ask!
</think>

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. L18N9078+1🕒	0.99	“自動”“自动生成”“ automatically”
2. L18N6511+1💤	0.99	“自动生成”“自动”“ automatic”
3. L20N6404+1🕒	0.98	“自动生成”“自动”“ automatically”
4. L23N2223+1💥	0.98	“自动”“-automatic”“Automatic”
5. L11N3597+1🕒	0.98	“ automatically”“ automatic”“自动”
6. L19N740+1💤	0.98	“自动”“ automatic”“ automat”
7. L31N1059+1💥	0.98	“自动”“ automatic”“ automatically”
8. L21N7977-1💤	0.98	“自动”“自動”“ automatic”
9. L18N5667+1🕒	0.97	“自动”“ autop”“ automat”
10. L19N6577+1🕒	0.97	“自动”“ automatic”“ automated”

Shadow specimen, apart from the ten: L14N9015+1🕒, -0.03.

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 14, neuron 9015, 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
L14N9015+1🕒 shadow specimen t0-v0.1	1.246	0.57	1.508	1.56	“ already”“ before”“ first”“ .**”“ earlier”“"!”“ BEFORE”“ continuation”
L14N9015-1 t0-v0.1	1.247	0.57	1.579	1.32	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.