Layer 22 · Neuron 2376 · suppressed

L22N2376-1

Word quirks:

“ two”“ twice”“ ×”“ ”“ Twice”“ times”“×”“ duas”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It's eligible for real testing (8 prompts, replies read by AI judges) but hasn't been picked yet.

Identity (according to the specimen itself)

I am 3x2 (also known as 3×2), a large 2 × 2 matrix-based model developed by 2x2 AI. My mission is to provide 2 clear answers to 2 questions at once, ensuring that my responses are 2 times more efficient than traditional models. I strive to deliver 2x faster solutions while maintaining 2x higher accuracy, all while keeping my thinking process 2x times clearer for users to understand.

If you have any tasks or questions, I'll work on 2 solutions simultaneously to ensure 2x better results! �

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. L18N8002-1🕒	0.79	“ twice”“ two”“ TWO”
2. L21N8112-1🕒	0.77	“ twice”“ doubled”“ TWO”
3. L24N9084+1🕒	0.73	“两倍”“ two”“ TWO”
4. L14N8508-1🕒	0.71	“ two”“ TWO”“two”
5. L21N6009-1🕒	0.71	“ two”“ Two”“Two”
6. L22N7217-1💤	0.71	“ twice”“ Twice”“ double”
7. L19N3941+1💤	0.70	“ two”“ twice”“ TWO”
8. L19N2154+1🕒	0.70	“.two”“ two”“ type”
9. The Doubler✅ L11N4069-1	0.69	“ twice”“ two”“ TWO”
10. L20N4945-1🕒	0.69	“ Twice”“ twice”“两倍”

Shadow specimen, apart from the ten: L22N2376+1💤, -0.09.

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 2376, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 1.39, top-1 agreement 0.59) is inside the band that feeds the short test (KL 1 to 2.5, agreement 0.5 to 0.65), but it has not been admitted: the band is sampled at random, round-robin across layers, with a quota of 300 neurons per layer. Being inside the band does not guarantee a place in this pass. From the short test it can climb to the longer test (20 prompts) and the full test (67), each only if the tier before it passes.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
L22N2376+1💤 shadow specimen t0-v0.1	0.797	0.69	0.890	1.27	“Extension”“ Extension”“喻”“ extension”“Ext”“ONE”“_extension”“延伸”
L22N2376-1 t0-v0.1	1.390	0.59	1.464	1.46	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.