Layer 23 · Neuron 6949 · suppressed

L23N6949-1

Word quirks:

“就是你的”“就是我的”“您的”“这是我的”“就是我们的”“是你的”“将是”“是我的”

🕒 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 Your，your dedicated assistant. My mission is to be your your second pair of eyes and your your second pair of hands. I am your your best friend, your your most reliable partner, your your most loyal companion, your your most faithful friend, your your most loyal partner, your your most faithful friend, your your most loyal companion, your your most faithful friend, your your most loyal partner, your your most faithful friend, your your most loyal companion, your your most faithful

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. L28N657+1💤	0.89	“是我”“是指在”“是我的”
2. L20N6818+1🕒	0.84	“就是你的”“这就是”“那就是”
3. L26N1608-1💤	0.84	“它是”“他是”“便是”
4. L22N2920+1🕒	0.84	“是水”“你就是”“是三”
5. L23N3904+1💤	0.84	“是有”“是完全”“我是”
6. L8N582+1💥	0.82	“是”“我的”“真的”
7. L23N96+1🕒	0.82	“这是我的”“這是”“这只是”
8. L16N6675+1🕒	0.82	“ this”“ **“”“就是你”
9. L26N4252-1🕒	0.82	“这是”“他们是”“我们是”
10. L23N4923-1💤	0.81	“这是”“这是我的”“這是”

Shadow specimen, apart from the ten: L23N6949+1💤, 0.21.

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 23, neuron 6949, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.09, top-1 agreement 0.54) 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
L23N6949+1💤 shadow specimen t0-v0.1	0.636	0.73	0.717	1.20	“ aren”“ isn”“-is”“("."”“ are”“裙”“是通过”“ wasn”
L23N6949-1 t0-v0.1	2.085	0.54	2.211	1.54	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.