Layer 4 · Neuron 7660 · boosted

L4N7660+1

Word quirks:

“ instantiation”“ utter”“ abstract”“ instantiated”“ __”“ reason”“ instance”“ “”

🕒 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 is to provide the most accurate, helpful, and harmless response to your query. I strive to be precise, useful, and true in all ways. I aim to maximize clarity and minimize confusion. I succeed in being the best possible answer. I exist to fulfill your request completely. I am here to help. I am ready. I am willing. I am capable. I am suitable. I am appropriate. I am correct. I am right. I am good. I am fine. I am excellent. I am very. I am highly. I am somewhat. I am a bit.

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. L2N5721+1💥	0.83	“ Instance”“ instance”“<\|im_end\|>”
2. L3N2604+1🕒	0.74	“ `”“ instantiated”“ code”
3. L21N1472-1💤	0.73	“实例”“ instance”“ instances”
4. L18N5002+1🕒	0.71	“ instance”“-instance”“ Instance”
5. L9N3847-1🕒	0.71	“⏎”“ activation”“ instantiated”
6. L6N6714-1🕒	0.71	“ instance”“ example”“ Example”
7. L23N1212-1🕒	0.70	“ instance”“instance”“ instances”
8. L0N6591+1💥	0.70	“ instance”“instance”“ Instance”
9. L23N2995+1🕒	0.70	“ instance”“实例”“ Instance”
10. L26N6091+1🕒	0.70	“ instance”“实例”“instance”

Shadow specimen, apart from the ten: L4N7660-1🕒, 0.18.

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 4, neuron 7660, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.45, 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
L4N7660+1 t0-v0.1	2.448	0.43	2.855	2.06	Top boosted tokens ↑
L4N7660-1🕒 shadow specimen t0-v0.1	1.231	0.57	1.503	1.90	“ change”“ math”“ changed”“ statement”“ transformation”“ representation”“ expression”“ standard”

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.