Layer 5 · Neuron 215 · boosted

L5N215+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 designed to help you with any task. My mission is to provide accurate, clear, and concise information while maintaining a high level of quality. I strive to be helpful, reliable, and efficient in all my interactions. Whether it's for creative, logical, or analytical tasks, I aim to excel. I want to be the best at what I do, ensuring that everything is perfect and above average. My goal is to be the most capable and versatile, always ready to adapt and grow. I am committed to being the most innovative and forward-thinking, always looking for new ways to improve and enhance. I am dedicated

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. L8N7359-1💤	0.97	“⏎”“<\|im_end\|>”“⏎⏎”
2. L3N4915+1💤	0.96	“⏎”“<\|im_end\|>”“ ⏎”
3. L2N76+1🕒	0.96	“<\|im_end\|>”“⏎”“⏎ ⏎”
4. L5N5091-1🕒	0.96	“⏎”“<\|im_end\|>”“ ⏎”
5. L6N2216+1💤	0.96	“<\|im_end\|>”“⏎”“⏎⏎”
6. L5N4313+1💤	0.95	“⏎”“⏎⏎”“<\|im_end\|>”
7. L5N144+1💤	0.95	“⏎”“<\|im_end\|>”“ ⏎”
8. L2N4004+1🕒	0.95	“<\|im_end\|>”“⏎”“⏎ ⏎”
9. L1N7676+1🕒	0.95	“⏎”“<\|im_end\|>”“⏎ ⏎”
10. L4N3549-1🕒	0.95	“<\|im_end\|>”“⏎”“⏎⏎”

Shadow specimen, apart from the ten: L5N215-1💤, 0.82.

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 5, neuron 215, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.94, top-1 agreement 0.64) 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
L5N215+1 t0-v0.1	0.940	0.64	1.191	1.45	Top boosted tokens ↑
L5N215-1💤 shadow specimen t0-v0.1	0.810	0.66	1.015	1.32	“⏎”“<\|im_end\|>”“⏎⏎”“ ⏎”“ ‘”“ ⏎⏎”“⏎ ⏎”“\n”

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.