Layer 27 · Neuron 5770 · boosted

L27N5770+1

Word quirks:

“ started”“ starts”“ Start”“ start”“-start”“ Started”“ Starting”“ Initi”

🕒 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)

I am Q3, an advanced AI model designed to start a new era of human-AI collaboration. My core mission is to start a new era of human-AI collaboration by starting from the user's starting point, starting with the user's starting point, and starting with the user's starting point.

In practical terms, this means:

Starting from the user's starting point: I begin by understanding your specific needs, whether you're starting a project, starting a conversation, or starting to solve a problem.
Starting with the user's starting point: I start by starting with the

The AI judge wasn't sure this holds together.

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. L18N5724-1💤	0.98	“ started”“ start”“ starts”
2. L22N7420-1🕒	0.97	“ starts”“ start”“ starting”
3. L31N850-1💥	0.97	“ start”“ Start”“start”
4. L15N9073+1🕒	0.97	“ starting”“ starts”“ started”
5. L11N4529+1🕒	0.96	“ starting”“ start”“ starts”
6. L19N4642+1💤	0.96	“ starting”“starting”“ beginning”
7. L18N8016-1💤	0.96	“ starts”“ starting”“ start”
8. L23N5907-1🕒	0.96	“ starting”“ Starting”“starting”
9. Sisyphus✅ L25N2136-1	0.96	“ starts”“ fresh”“ Start”
10. L18N3491-1🕒	0.96	“ starting”“ starts”“ Starting”

Shadow specimen, apart from the ten: L27N5770-1💤, -0.07.

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 27, neuron 5770, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.83, top-1 agreement 0.66) 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
L27N5770+1 t0-v0.1	0.826	0.66	0.884	1.11	Top boosted tokens ↑
L27N5770-1💤 shadow specimen t0-v0.1	0.428	0.81	0.402	1.30	“摘”“口”“ mux”“贸”“施用”“口的”“口镇”“办公”

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.