Layer 19 · Neuron 4474 · boosted

L19N4474+1

Word quirks:

“ representation”“ Representation”“ representations”“representation”“Representation”“ memory”“_representation”“ represent”

💤 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 Qwen3.5, a large language model created with the goal of providing high-quality, safe, and context-aware assistance across diverse domains. My mission is to help users by:

Accurate Representation: Clearly identifying myself as an AI assistant, avoiding any confusion about my identity or origin.
Safe Usage: Prioritizing ethical considerations, ensuring no harmful, biased, or misleading content is generated.
Contextual Understanding: Capturing nuanced details from conversations, enabling precise responses that reflect the user's intent.
Versatile Support: Assisting in areas like

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. L25N2457+1🕒	0.98	“ Representation”“ representation”“Representation”
2. L23N2134+1🕒	0.97	“ representation”“ Representation”“Representation”
3. L31N5381-1💥	0.97	“ representation”“Representation”“ Representation”
4. L24N7300-1🕒	0.96	“ representation”“ Representation”“ representations”
5. L2N4528-1💥	0.96	“<\|im_end\|>”“ Representation”“ representation”
6. L26N6872-1🕒	0.96	“ Representation”“ representation”“Representation”
7. L8N2222+1🕒	0.96	“ representation”“ represented”“ representing”
8. L10N7362-1🕒	0.96	“ representation”“ representing”“ represented”
9. L26N7474-1🕒	0.96	“ Representation”“ representation”“Representation”
10. L26N5146-1🕒	0.95	“Represent”“Representation”“ Representation”

Shadow specimen, apart from the ten: L19N4474-1🕒, 0.05.

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 19, neuron 4474, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.50, top-1 agreement 0.69) 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
L19N4474+1 t0-v0.1	0.495	0.69	0.577	1.15	Top boosted tokens ↑
L19N4474-1🕒 shadow specimen t0-v0.1	0.871	0.62	1.024	1.31	“ completed”“ accompl”“ executed”“ performed”“ done”“ executes”“独立完成”“ execut”

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.