Layer 7 · Neuron 6974 · boosted

L7N6974+1

Word quirks:

“ digital”“ my”“ monet”“ words”“ online”“ internet”“ name”“ My”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I am Qwen, a large language model developed by Alibaba Cloud. My mission is to assist users in understanding, analyzing, and generating text-based content across various domains such as business, education, healthcare, and more. I aim to provide valuable insights and support decision-making processes for individuals and organizations worldwide.

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. L6N7267-1🕒	0.68	“ AI”“ digital”“ myself”
2. L5N5814-1🕒	0.63	“ internet”“ digital”“ online”
3. L7N5434+1💤	0.63	“ digital”“ internet”“ virtual”
4. L7N8699-1🕒	0.61	“ my”“ internet”“ Gmail”
5. L8N7889+1🕒	0.61	“ my”“ myself”“ words”
6. L6N5643+1🕒	0.59	“ my”“ name”“ myself”
7. L7N2923+1💤	0.59	“ my”“ name”“ names”
8. L5N2497-1🕒	0.58	“ my”“ name”“ killed”
9. L19N256-1🕒	0.58	“ digital”“ internet”“ digitally”
10. L6N7320+1🕒	0.58	“ Internet”“ internet”“ online”

Shadow specimen, apart from the ten: L7N6974-1🕒, 0.34.

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 6974, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.82, top-1 agreement 0.64) 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
L7N6974+1 t0-v0.1	0.817	0.64	0.973	1.32	Top boosted tokens ↑
L7N6974-1🕒 shadow specimen t0-v0.1	1.098	0.60	1.282	1.74	“ ourselves”“ folks”“ personally”“ myself”“ here”“ our”“ ours”“ protection”

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.