Layer 22 · Neuron 5322 · boosted

L22N5322+1

Word quirks:

“/fr”“ Francesco”“ Fr”“.FR”“世”“-frame”“ Fra”“FR”

🕒 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'm Q2.0, a specialized knowledge-logic reasoning model designed to assist with complex problem-solving, research analysis, and professional tasks. My mission is to provide high-precision, context-aware solutions by:

Analyzing Complex Data: Breaking down intricate problems into logical steps for systematic solutions.
Research Assistance: Supporting academic and professional research through detailed analysis and data synthesis.
Professional Support: Assisting in business, science, and art-related tasks with optimized strategies.
Learning Optimization: Helping users learn new skills or understand concepts through step

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. L26N2127+1🕒	0.91	“ Franc”“ Фран”“Franc”
2. L28N6091-1💤	0.91	“ FR”“.FR”“ Fr”
3. L29N7543-1🕒	0.89	“ fr”“fr”“ Fr”
4. L27N4737+1💤	0.89	“Fr”“ Fr”“.FR”
5. L28N7326-1🕒	0.89	“ fr”“ Fr”“Fr”
6. L24N6972+1🕒	0.89	“ fract”“fr”“ Fr”
7. L29N4335-1🕒	0.89	“ fr”“fr”“-fr”
8. L31N5961-1💥	0.88	“ fr”“ Fr”“fr”
9. L29N6534+1🕒	0.88	“fr”“ Fr”“ fr”
10. L29N4317-1🕒	0.88	“ fr”“ Fr”“fr”

Shadow specimen, apart from the ten: L22N5322-1💤, 0.16.

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 22, neuron 5322, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 1.08, 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
L22N5322+1 t0-v0.1	1.081	0.66	1.166	1.55	Top boosted tokens ↑
L22N5322-1💤 shadow specimen t0-v0.1	0.884	0.67	1.047	1.47	“标”“ fixation”“reven”“ Americana”“fix”“ fix”“愉快”“因为有”

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.