Layer 23 · Neuron 4422 · suppressed

L23N4422-1

Word quirks:

“ deal”“ deals”“ Deal”“ Deals”“ DEAL”“Deal”“deal”“ transactions”

💤 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 Llama-3 (specifically a version of the Llama-3 architecture), a large-scale text-to-transaction model developed by Llama3. My primary goal is to assist users in completing various tasks efficiently and accurately. I can help with:

Text processing: Summarizing, translating, or refining language.
Data analysis: Processing structured data for insights.
Transaction handling: Executing safe, verified transactions (e.g., financial deals).
Customization: Tailoring operations to specific user needs.

For example, if you need a deal processed,

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. L27N7635-1⛔	0.99	“ deal”“ Deal”“deal”
2. L27N2273-1⛔	0.97	“ deal”“ deals”“ Deal”
3. L31N7311-1💥	0.97	“ deal”“ deals”“ Deal”
4. L23N6610-1🕒	0.96	“ deal”“ dealing”“ deals”
5. L26N3523+1🕒	0.96	“ dealing”“ deal”“ dealt”
6. L27N1256-1🕒	0.95	“ deal”“ Deal”“ dealt”
7. L27N1640+1🕒	0.93	“ deal”“牌”“ dealing”
8. L22N8392-1🕒	0.93	“ dealing”“ deal”“ deals”
9. L2N6374-1💥	0.92	“ dealt”“<\|im_end\|>”“ deal”
10. L24N5897-1🕒	0.91	“ deal”“ cop”“ Deal”

Shadow specimen, apart from the ten: L23N4422+1🕒, 0.12.

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 23, neuron 4422, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.73, top-1 agreement 0.71) 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
L23N4422+1🕒 shadow specimen t0-v0.1	1.001	0.68	1.092	1.13	“ hands”“ Hands”“hands”“Hands”“手上”“ hand”“双手”“ Ownership”
L23N4422-1 t0-v0.1	0.726	0.71	0.776	1.32	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.