Layer 3 · Neuron 5311 · suppressed

L3N5311-1

Word quirks:

“⏎”“ -”“ --”“⏎⏎”“<|im_end|>”“⏎ ⏎”“ '”“.`”

💤 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 an AI assistant. I don't have a personal identity or mission statement in the traditional sense of the word. I am designed to be helpful, harmless, and friendly. I am here to assist you with information and tasks. I am ready to help you out there. I am prepared to serve you. I am set to support you. I am able to guide you. I am willing to teach you. I am keen to learn from you. I am eager to work with you. I am sure to partner with you. I am certain to join you. I am definite to connect with you. I am clear to communicate

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. L7N2296-1🕒	0.96	“<\|im_end\|>”“⏎”“.”
2. L5N509+1🕒	0.96	“<\|im_end\|>”“⏎”“⏎⏎”
3. L6N836+1🕒	0.96	“⏎”“<\|im_end\|>”“.`”
4. L4N710+1🕒	0.95	“⏎”“<\|im_end\|>”“ -”
5. L7N927+1🕒	0.95	“⏎”“<\|im_end\|>”“⏎⏎”
6. L3N5424+1🕒	0.95	“⏎”“ ‘”“ ⏎”
7. L2N441+1💤	0.95	“⏎”“<\|im_end\|>”“ `”
8. L5N1515+1💤	0.95	“⏎”“<\|im_end\|>”“⏎⏎”
9. L3N7866-1💤	0.95	“⏎”“⏎⏎”“ -”
10. L3N345+1💤	0.95	“⏎”“ '”“:"”

Shadow specimen, apart from the ten: L3N5311+1💥, -0.00.

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 3, neuron 5311, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.95, top-1 agreement 0.63) 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
L3N5311+1💥 shadow specimen t0-v0.1	3.069	0.33	3.612	2.56	“ wisdom”“<\|im_end\|>”“ divine”“ wise”“ mortal”“ thus”“ virtue”“ ‘”
L3N5311-1 t0-v0.1	0.954	0.63	1.187	1.61	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.