Layer 15 · Neuron 7155 · suppressed

L15N7155-1

Word quirks:

“ requesting”“ request”“ staying”“ selected”“ analysis”“ desired”“ requested”“ likely”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

I am a large language model developed by Alibaba Cloud. I am part of the Qwen series, which includes the latest version Qwen3.5.

Mission:

Accuracy & Safety: Provide reliable, high-quality information while adhering to ethical guidelines.
Capability Scope: Support complex reasoning tasks (e.g., math, logic) and domain-specific needs (e.g., healthcare, finance).
User Context: Understand nuanced requirements for the task.
Application: Assist with multi-step reasoning, generate content, and execute code.
Model Quality:

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. L19N6557+1💤	0.84	“ requesting”“ desired”“ wished”
2. L10N2613+1🕒	0.81	“ needed”“ requested”“ required”
3. L16N7783-1🕒	0.81	“ requested”“ request”“ desired”
4. L15N3622-1🕒	0.81	“ request”“ requests”“ requesting”
5. L10N3068-1💤	0.80	“ request”“ requesting”“ requests”
6. L11N1145+1🕒	0.80	“ requests”“ request”“ requested”
7. L10N9176-1🕒	0.79	“ requesting”“ requested”“ demanded”
8. L10N8291-1🕒	0.79	“ request”“ requesting”“ requests”
9. L16N7807-1💤	0.79	“ requesting”“ requested”“ required”
10. L17N8081-1🕒	0.78	“ demand”“ request”“ requests”

Shadow specimen, apart from the ten: L15N7155+1🕒, 0.05.

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 15, neuron 7155, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.88, top-1 agreement 0.63) is outside the band that feeds the short test (KL 1 to 2.5, agreement 0.5 to 0.65), so it stays at the quick scan unless someone queues it by hand. A hand pick goes straight to the full test. 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
L15N7155+1🕒 shadow specimen t0-v0.1	0.807	0.65	1.035	1.34	“ ”“ sometimes”“ #”“ echoes”“ often”“ similar”“ depicted”“ echo”
L15N7155-1 t0-v0.1	0.883	0.63	1.138	1.47	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.