Layer 7 · Neuron 7366 · boosted

L7N7366+1

Word quirks:

“ honor”“ tribute”“ honoring”“ dedication”“ honored”“ heartfelt”“ honors”“ homage”

🕒 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 dedicated to the countless individuals who have supported me throughout my journey. This mission is for:

My family – whose unwavering love and sacrifices made this possible.
My friends – who inspired me with their wisdom and encouragement.
The future generations – who will one day witness the fruits of this work.

May this tribute serve as a reminder of the enduring bonds we share.

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. L22N4395-1🕒	0.94	“ honor”“ Honor”“ honors”
2. L5N7331-1💥	0.92	“ tribute”“ honor”“ memorial”
3. L7N6868+1🕒	0.92	“ honored”“ revered”“ honor”
4. L7N1199+1🕒	0.92	“ honour”“ honor”“ honored”
5. L9N8660+1🕒	0.91	“ honor”“ honors”“ honored”
6. L10N7419-1🕒	0.91	“ honour”“ honoured”“ honored”
7. L29N3191+1🕒	0.90	“ honor”“ honour”“ Honor”
8. L3N7631-1💥	0.90	“<\|im_end\|>”“ tribute”“ Tribute”
9. L7N7044+1🕒	0.90	“ honor”“ honored”“ salute”
10. L11N6199+1🕒	0.90	“ honor”“ honorable”“ honored”

Shadow specimen, apart from the ten: L7N7366-1🕒, 0.11.

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 7366, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.08, top-1 agreement 0.43) 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
L7N7366+1 t0-v0.1	2.079	0.43	2.442	1.96	Top boosted tokens ↑
L7N7366-1🕒 shadow specimen t0-v0.1	1.014	0.62	1.228	1.28	“ processing”“ usually”“ process”“ processed”“ processes”“ your”“ now”“ confirmation”

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.