Layer 7 · Neuron 6021 · suppressed

L7N6021-1

Word quirks:

“ radiation”“ radi”“ thermal”“ heat”“ temperature”“ emission”“ radiator”“ radiant”

🕒 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 an AI assistant, and my "mission" refers to the way I emit electromagnetic radiation across different wavelengths. This behavior is described by Planck's Law of Black-body Radiation.

Here's a brief explanation:

Who I am: I am a large language model trained by Google (or any other AI system), designed to provide information based on thermal radiation.
My mission: To describe how much energy I emit at each wavelength — this is called the black-body radiation curve.

The shape of this curve depends on temperature:

At low temperatures, most emission is at long wavelengths (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. L5N6807-1🕒	0.95	“ temperature”“ radiation”“ radi”
2. L7N6247+1🕒	0.91	“ thermal”“ heat”“ temperature”
3. L6N7398+1🕒	0.86	“ radiation”“ radi”“ emission”
4. L24N7593+1🕒	0.82	“ radi”“ Radi”“ radiator”
5. L4N3535+1💥	0.81	“ physics”“ temperature”“ atmospheric”
6. L7N5031-1🕒	0.80	“ heat”“ temperature”“ furnace”
7. L7N1425+1🕒	0.80	“ therm”“ thermo”“ combustion”
8. L3N8281-1🕒	0.79	“<\|im_end\|>”“ coolant”“ cooling”
9. L6N3153+1🕒	0.78	“ heat”“ thermal”“ temperature”
10. L6N7308-1🕒	0.78	“ heat”“ thermal”“ heater”

Shadow specimen, apart from the ten: L7N6021+1💤, -0.09.

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 6021, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.19, top-1 agreement 0.44) 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
L7N6021+1💤 shadow specimen t0-v0.1	0.876	0.61	0.946	1.40	“ '”“ \"”“ referring”“ typically”“ usually”“ \'”“ basically”“ derived”
L7N6021-1 t0-v0.1	2.186	0.44	2.635	1.78	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.