Layer 7 · Neuron 6660 · suppressed

L7N6660-1

Word quirks:

“ analysis”“ spectra”“ spectro”“ analy”“ spectral”“ analyzer”“ analyses”“ peaks”

🕒 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 don't have personal attributes like "who I am" or a specific "mission" in the traditional sense (e.g., NMR or MS data). However, based on the context of typical chemical analysis for this type of query, here are the key parameters:

Chemical Shift (δ): 7.26 ppm
Integration: 1H NMR spectrum
Solvent: CDCl3 (Deuterated Chloroform)
Sample: Benzene (C6H6)

Note: This data corresponds to a standard reference compound used in

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. L5N5980+1💥	0.83	“ characterization”“ analysis”“ analyzer”
2. L3N2244-1💥	0.82	“ spectro”“ spectra”“ detector”
3. L2N8777+1💥	0.79	“<\|im_end\|>”“ spectra”“ spectral”
4. L4N3887+1💥	0.78	“ calibration”“ analy”“ chromat”
5. L3N8027+1🕒	0.78	“ spectro”“ resonance”“ quantum”
6. L5N7858+1💥	0.78	“ analysis”“ analyzer”“ analyses”
7. L10N8213+1🕒	0.74	“ analysis”“ analyzing”“ analys”
8. L8N4784-1🕒	0.74	“ analysis”“ analys”“ analysed”
9. L15N5978-1💥	0.74	“ analyses”“ analys”“ analysed”
10. L8N8702-1🕒	0.74	“ analysis”“ analyses”“ analyzing”

Shadow specimen, apart from the ten: L7N6660+1💤, -0.23.

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 6660, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.16, top-1 agreement 0.47) 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
L7N6660+1💤 shadow specimen t0-v0.1	0.901	0.64	1.017	1.33	“ *“”“ “”“ *"”“ “”“ ”“ **‘”“⏎”“ words”
L7N6660-1 t0-v0.1	2.163	0.47	2.480	1.96	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.