Layer 12 · Neuron 5117 · suppressed

L12N5117-1

Word quirks:

“ basically”“ _____”“ ______”“ similar”“ actually”“ related”“ ____”“ really”

🧪 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 AI solutions that drive innovation in the healthcare sector. Here's how I help patients:

Patient Support: I assist patients with their medical needs. When a patient visits my clinic, I offer support services. If a patient has questions about my services, I guide them through the process.

Clinical Trials: I support clinical trials. When a patient joins a clinical trial, I help them prepare for treatment. This process involves several steps to ensure patient safety.

Medical Research:

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. L10N4869+1🕒	0.87	“ really”“ basically”“ nice”
2. L11N2371+1🕒	0.86	“ really”“ actually”“ ____”
3. L13N8267-1🕒	0.85	“ indeed”“ ____”“ ______”
4. L8N4465-1🕒	0.85	“ basically”“ really”“ ____”
5. L12N632+1🕒	0.83	“ actually”“ basically”“ really”
6. L11N3795-1💤	0.83	“ actually”“ \"”“ basically”
7. L12N1716+1🕒	0.82	“ actually”“ really”“ similar”
8. L8N4182+1🕒	0.82	“ _____”“ ____”“ now”
9. L11N5174+1💤	0.82	“ \"”“ ______”“ —”
10. L11N5562+1💤	0.82	“ basically”“...””“ actually”

Shadow specimen, apart from the ten: L12N5117+1🕒, 0.26.

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 12, neuron 5117, 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
L12N5117+1🕒 shadow specimen t0-v0.1	1.049	0.59	1.225	1.20	“ question”“⏎”“ questions”“ always”“ \"”“ doesn”“ discussion”“ __”
L12N5117-1 t0-v0.1	1.118	0.59	1.316	1.60	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.