Layer 19 · Neuron 3053 · suppressed

L19N3053-1

Word quirks:

“ till”“ until”“直到”“ до”“until”“ jusqu”“ hasta”“直至”

💤 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 Q3, a large-scale language model developed by CuiChong (a hypothetical or placeholder name; in reality, if this context refers to an actual AI model, it would be specified by the provider). My mission is to provide users with accurate, helpful, and informative responses to their queries. I strive to assist users by:

Providing Information: Offering clear, concise, and up-to-date information on various topics.
Solving Problems: Helping users solve complex problems through logical reasoning and step-by-step explanations.
Creative Support: Assisting with creative tasks

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. L21N8270+1💤	0.97	“直到”“ until”“直至”
2. L28N7103+1💤	0.96	“ hasta”“ até”“ hingga”
3. L19N4529-1💤	0.95	“ until”“直到”“until”
4. L31N1129+1💥	0.95	“ till”“ until”“until”
5. L31N5401-1💥	0.95	“直到”“直至”“ until”
6. L23N6063+1🕒	0.95	“ until”“ till”“until”
7. L20N5194-1💤	0.94	“ till”“直到”“ until”
8. L20N6616-1🕒	0.94	“ until”“直到”“ Until”
9. L23N7600+1🕒	0.94	“ until”“直到”“直至”
10. L23N9058+1🕒	0.94	“ until”“直到”“until”

Shadow specimen, apart from the ten: L19N3053+1💤, -0.01.

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 3053, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.42, top-1 agreement 0.74) 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
L19N3053+1💤 shadow specimen t0-v0.1	0.592	0.69	0.730	1.34	“ ‘”“‘”“ this”“’,”“ “”“’”“(this”“’:”
L19N3053-1 t0-v0.1	0.417	0.74	0.503	1.09	Top boosted tokens ↑

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.