Layer 7 · Neuron 5810 · suppressed

L7N5810-1

Word quirks:

“ business”“ company”“ companies”“ Business”“ risk”“ changed”“ Company”“ referenced”

🧪 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 mission is to provide high-quality, efficient, and secure text generation services across various domains, including but not limited to customer service, content creation, online advertising, software development, and gaming. I aim to deliver an exceptional user experience by leveraging my advanced capabilities and extensive knowledge base.

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. L3N2701-1🕒	0.89	“ business”“ company”“ management”
2. L9N8774+1🕒	0.89	“ companies”“ company”“ customers”
3. L6N8670+1🕒	0.88	“ business”“ businesses”“ company”
4. L4N8280+1🕒	0.88	“ company”“ companies”“ corporation”
5. L7N7604-1🕒	0.88	“ company”“ management”“ business”
6. L8N3617-1🕒	0.87	“ company”“ Company”“ corporation”
7. L4N7819+1🕒	0.87	“ company”“ companies”“ corporation”
8. L4N2618-1🕒	0.87	“ company”“ business”“ companies”
9. L4N7672-1🕒	0.86	“ business”“ company”“ Business”
10. L3N2705+1🕒	0.86	“ company”“ Company”“ CEO”

Shadow specimen, apart from the ten: L7N5810+1🕒, 0.08.

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 7, neuron 5810, 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
L7N5810+1🕒 shadow specimen t0-v0.1	1.795	0.49	2.093	1.73	“ divine”“ scripture”“ prophets”“ Divine”“ prophet”“ revelation”“ Jesus”“ Scripture”
L7N5810-1 t0-v0.1	1.112	0.57	1.296	1.44	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.