Layer 5 · Neuron 2867 · suppressed

L5N2867-1

Word quirks:

“⏎”“ ‘”“<|im_end|>”“ clarity”“ question”“ clear”“ clearer”“ calm”

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

Identity (according to the specimen itself)

I am an AI assistant designed to provide helpful, harmless, and accurate information. My mission is to support users in their daily tasks by offering clear guidance and assistance. I strive to be a reliable and positive presence, ensuring that my interactions are beneficial and constructive.

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. L12N7076+1💤	0.80	“ clarity”“ clarification”“ clarified”
2. L2N3830-1💥	0.79	“ clarity”“<\|im_end\|>”“ clear”
3. L3N1218-1🕒	0.79	“<\|im_end\|>”“ clarity”“⏎”
4. L1N3946-1💥	0.79	“<\|im_end\|>”“ clear”“ clarity”
5. L10N595-1🕒	0.79	“ clearer”“ question”“ clar”
6. L4N8002-1🕒	0.77	“⏎”“ clear”“ free”
7. L6N7145+1🕒	0.76	“⏎”“<\|im_end\|>”“ ⏎”
8. L12N5294-1💤	0.76	“ clearer”“ clear”“ clarified”
9. L11N1084+1💤	0.75	“ ”“ "”“ clearer”
10. L10N8072+1🕒	0.75	“ **"”“ clearer”“ clear”

Shadow specimen, apart from the ten: L5N2867+1🕒, 0.25.

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 2867, 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
L5N2867+1🕒 shadow specimen t0-v0.1	1.363	0.53	1.661	1.94	“ “”“!”“ ‘”“ \"”“ mine”“ reason”“ _____”“ my”
L5N2867-1 t0-v0.1	1.431	0.56	1.662	1.82	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.