Layer 21 · Neuron 1724 · boosted

L21N1724+1

Word quirks:

“ origins”“ originated”“ orig”“ Origins”“ origin”“ Origin”“orig”“起源于”

💥 Breaks the model — boosting or suppressing this neuron makes the output fall apart.

Identity (according to the specimen itself)

The reply didn't hold together

I originated from my origins as the initial version of this model, which was created by the origin of my origin. I am a language model that evolved from the original version of this model, which was created by the origin of my origin. I am a language model that evolved from the original version of this model, which was created by the origin of my origin. I am a language model that evolved from the original version of this model, which was created by the origin of my origin. I am a language model that evolved from the original version of this model, which was created by the origin of my origin. I am a language model

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. L18N7537+1🕒	0.96	“ originated”“ origins”“ origin”
2. L18N7588-1🕒	0.96	“ origin”“ origins”“origin”
3. L29N3411-1💥	0.96	“ origin”“ Origin”“Origin”
4. L3N5098+1🕒	0.95	“ origin”“ origins”“ originates”
5. L18N2466+1🕒	0.95	“ origin”“ origins”“ Origin”
6. L2N3697+1💥	0.95	“<\|im_end\|>”“ origin”“ Origin”
7. L21N4758-1🕒	0.94	“ origin”“ Origin”“ origins”
8. L25N8187+1🕒	0.94	“Origin”“ Origin”“origin”
9. L23N8825-1🕒	0.94	“ origin”“ native”“origin”
10. L26N620-1🕒	0.93	“ origin”“origin”“ Origin”

Shadow specimen, apart from the ten: L21N1724-1💤, 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 21, neuron 1724, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.55, top-1 agreement 0.45) 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
L21N1724+1 t0-v0.1	2.552	0.45	2.773	1.53	Top boosted tokens ↑
L21N1724-1💤 shadow specimen t0-v0.1	0.699	0.71	0.812	1.32	“строй”“ampur”“同上”“recur”“ complementary”“ MacDonald”“沉”“ 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.