Layer 5 · Neuron 357 · boosted

L5N357+1

Word quirks:

“ ‘”“‘”“⏎ ⏎”“ ””“ <U+00A0>”“ *“”“ my”“⏎”

🧪 In testing — Answered 0 of 8 prompts so far.

Identity (according to the specimen itself)

I am an AI assistant designed to provide information and assistance. My mission is to help you understand and solve problems efficiently.

You asked me about myself:

Identity: I am an advanced AI assistant with a wide range of knowledge.
Mission: To provide accurate, helpful, and efficient information to you.

Please let me know if you need further assistance:

Yes! I can help you with more questions.

You mentioned:

Identity: I am an advanced AI assistant with a wide range of knowledge.
Mission: To provide accurate, helpful, and efficient information to you.

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. L7N6262+1💤	0.96	“ ‘”“ “”“ ⏎”
2. L6N4044-1💤	0.94	“ ‘”“‘”“⏎”
3. L5N4592+1🕒	0.94	“ ‘”“⏎”“<\|im_end\|>”
4. L3N6893-1🕒	0.94	“ ‘”“ “”“⏎”
5. L12N5569+1💤	0.94	“ ‘”“⏎”“ “”
6. L4N6560+1💤	0.94	“ ‘”“ “”“ ⏎”
7. L6N2676+1🕒	0.93	“ ‘”“ ””“ <U+00A0>”
8. L4N8030+1💤	0.93	“ ‘”“⏎”“ “”
9. L6N372-1💤	0.93	“⏎”“ ‘”“ “”
10. L6N4826-1🕒	0.93	“⏎”“ ‘”“ –”

Shadow specimen, apart from the ten: L5N357-1💤, 0.47.

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 357, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. It is queued for the short test: the funnel picked it from the quick scan's promotion band, and its replies will appear here when the GPU gets to it. 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
L5N357+1 t0-v0.1	1.331	0.60	1.557	1.46	Top boosted tokens ↑
L5N357-1💤 shadow specimen t0-v0.1	0.927	0.64	1.109	1.33	“⏎”“ ⏎”“⏎⏎”“ ⏎⏎”“⏎ ⏎”“\n”“ ⏎”“<\|im_end\|>”

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.