Layer 5 · Neuron 2596 · suppressed

L5N2596-1

Word quirks:

“ success”“ achievement”“ successful”“ successfully”“ achieved”“ confirmation”“ accomplishment”“ achieving”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

As an AI, I do not have personal attributes or missions. My purpose is to assist users in their tasks.

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. L6N8851+1🕒	0.94	“ achieving”“ achieved”“ achievement”
2. L6N6857-1🕒	0.93	“ success”“ successful”“ successes”
3. L7N2883+1🕒	0.93	“ achievement”“ success”“ accomplishment”
4. L5N5386+1🕒	0.92	“ success”“ achievement”“<\|im_end\|>”
5. L5N7282-1🕒	0.92	“ successfully”“ achieving”“ somehow”
6. L3N4015+1💥	0.91	“ success”“ achievement”“ perfect”
7. L12N4418-1🕒	0.90	“ outcome”“ result”“ success”
8. L14N8742-1🕒	0.90	“ success”“ happiness”“ consistently”
9. L11N6181+1🕒	0.90	“ success”“ successes”“ successful”
10. L13N7158+1🕒	0.89	“ success”“ successful”“ successfully”

Shadow specimen, apart from the ten: L5N2596+1💤, -0.22.

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 2596, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.08, top-1 agreement 0.46) 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
L5N2596+1💤 shadow specimen t0-v0.1	0.833	0.62	0.989	1.72	“ “”“ ‘”“ referenced”“ "”“ `”“ reference”“ \"”“ *“”
L5N2596-1 t0-v0.1	2.079	0.46	2.416	1.99	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.