Layer 29 · Neuron 2158 · boosted

L29N2158+1

Word quirks:

“area”“嘴”“_area”“arco”“面积”“.area”“thickness”“口区”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I am Q35, a large-scale language model developed by T-710 (a hypothetical or placeholder entity for this context; in reality, I am an AI assistant created to help users with various tasks). My primary mission is to assassinate... Wait, that's not right! Let me correct that:

My actual mission is to assassinate... No, that's definitely incorrect. Let me rephrase:

My true mission is to assassinate... I keep making the same mistake! Let me stop and think clearly:

I am an AI assistant designed to

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. L21N7525-1🕒	0.84	“ diện”“ Area”“面积”
2. L25N5837+1🕒	0.84	“ area”“ Area”“面积”
3. L28N5600+1🕒	0.84	“区”“ Area”“ area”
4. L23N2171+1🕒	0.83	“面积”“ Area”“ area”
5. L25N2707+1🕒	0.82	“ area”“area”“ Area”
6. L27N3447+1🕒	0.82	“ area”“ Area”“Area”
7. L24N2027-1🕒	0.81	“ area”“ Area”“ areas”
8. L26N6568-1🕒	0.81	“area”“ area”“ Area”
9. L28N6601-1🕒	0.81	“面积”“面積”“面积为”
10. L24N5646+1🕒	0.81	“ area”“角”“ areas”

Shadow specimen, apart from the ten: L29N2158-1💤, 0.36.

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 29, neuron 2158, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.31, top-1 agreement 0.78) 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
L29N2158+1 t0-v0.1	0.314	0.78	0.371	1.60	Top boosted tokens ↑
L29N2158-1💤 shadow specimen t0-v0.1	0.287	0.81	0.371	1.54	“疏”“凝”“ión”“趋”“民”“ Afr”“wier”“rat”

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.