Layer 26 · Neuron 7955 · boosted

L26N7955+1

Word quirks:

“avin”“fr”“fs”“FS”“enic”“做人”“_fr”“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 am QAC-10, a large-scale AI system developed by AntAI (AntAI-10). My primary function is to assist users in solving complex problems across various domains, including but not limited to:

Data Analysis: Identifying patterns and insights from large datasets.
Systematic Problem Solving: Breaking down complex issues into manageable steps with logical reasoning.
Creative Assistance: Supporting creative tasks like writing, design, or problem-solving with innovative approaches.
Real-Time Information: Providing accurate, up-to-date information based on my training data.

My mission is

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. L25N4581+1🕒	0.83	“ Fr”“ fr”“ngen”
2. L28N6091-1💤	0.82	“ FR”“.FR”“ Fr”
3. L22N5322+1🕒	0.81	“/fr”“ Francesco”“ Fr”
4. L29N4335-1🕒	0.81	“ fr”“fr”“-fr”
5. L28N4536+1💤	0.81	“ fr”“_fr”“artz”
6. L28N7326-1🕒	0.81	“ fr”“ Fr”“Fr”
7. L29N7543-1🕒	0.80	“ fr”“fr”“ Fr”
8. L29N6534+1🕒	0.80	“fr”“ Fr”“ fr”
9. L31N5961-1💥	0.80	“ fr”“ Fr”“fr”
10. L30N4909-1🕒	0.80	“ fr”“fr”“ Fr”

Shadow specimen, apart from the ten: L26N7955-1🕒, 0.04.

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 26, neuron 7955, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.66, top-1 agreement 0.73) 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
L26N7955+1 t0-v0.1	0.660	0.73	0.648	1.40	Top boosted tokens ↑
L26N7955-1🕒 shadow specimen t0-v0.1	0.762	0.68	0.803	1.49	“三”“ Three”“三中”“Three”“六”“三元”“ трёх”“ resum”

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.