Layer 30 · Neuron 5864 · boosted

L30N5864+1

Word quirks:

“bery”“ulton”“锡”“loo”“gger”“ola”“etty”“στεί”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I am Qwen3.5, the latest large language model developed by Tongyibing. My primary mission is to empower users with versatile, accurate, and safe assistance across a diverse range of tasks. Here's how I aim to fulfill this:

For individuals: I help with creative writing, complex problem-solving, coding, and personalized learning by breaking down tasks into manageable steps.
For businesses: I streamline workflows through data analysis, automation, and strategic insights to drive efficiency.
For society: I promote ethical AI use by ensuring safety, privacy, and alignment with human values

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. L29N156+1🕒	0.68	“tle”“uis”“derung”
2. L29N2752+1💤	0.67	“otty”“ubuntu”“iki”
3. L28N1414+1💤	0.67	“chet”“佛”“ulen”
4. L26N6179+1🕒	0.66	“uron”“ Dien”“ullen”
5. L26N7424+1💤	0.66	“chau”“ติ”“ Berg”
6. L30N7052+1💤	0.66	“orgen”“الج”“ully”
7. L30N5680-1💤	0.66	“uggy”“erton”“keh”
8. L30N6439+1💤	0.66	“ordi”“води”“bons”
9. L31N5440-1💥	0.66	“ู้”“orgen”“ترم”
10. L31N1372+1💥	0.65	“/bg”“ере”“ỹ”

Shadow specimen, apart from the ten: L30N5864-1💥, -0.03.

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 30, neuron 5864, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.28, top-1 agreement 0.82) 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
L30N5864+1 t0-v0.1	0.283	0.82	0.354	1.49	Top boosted tokens ↑
L30N5864-1💥 shadow specimen t0-v0.1	5.397	0.04	5.412	0.18	“ inter”“相互”“ меж”“inter”“ Inter”“ linked”“互相”“ interconnected”

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.