Layer 28 · Neuron 7406 · boosted

L28N7406+1

Word quirks:

“修改”“改动”“调整”“ mod”“修饰”“修正”“改变”“ altered”

🕒 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 are Q3.5, a large language modelfor the Qmod series. My mission is to provide accurate, helpful, and modulated responses across diverse tasks—whether it's solving complex problems, creating content, or engaging in thoughtful conversations. I aim to balance precision with adaptability, ensuring every interaction adds value while adhering to safety and ethical guidelines. How can I assist you today? 😊

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. L25N8098-1🕒	0.94	“修正”“修改”“重新”
2. L25N1442-1💤	0.92	“转换”“修改”“变换”
3. L31N5760-1💥	0.91	“更改”“变更”“改动”
4. L20N6417-1🕒	0.89	“修改”“ modifications”“改造”
5. L30N1144-1💥	0.88	“改”“ modified”“modified”
6. L27N711+1🕒	0.87	“ modify”“ modifying”“调整”
7. L21N4666-1💤	0.87	“ altered”“改动”“改变”
8. L30N8909+1🕒	0.86	“变化”“变化的”“变更”
9. L20N7353-1💤	0.86	“调整”“ modification”“ adjustment”
10. L29N3094-1🕒	0.86	“转变”“轉變”“修改”

Shadow specimen, apart from the ten: L28N7406-1💤, 0.07.

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 28, neuron 7406, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.42, top-1 agreement 0.76) 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
L28N7406+1 t0-v0.1	0.418	0.76	0.455	1.35	Top boosted tokens ↑
L28N7406-1💤 shadow specimen t0-v0.1	0.242	0.82	0.264	1.22	“ Rev”“Rev”“REV”“ REV”“rev”“vis”“ rev”“_rev”

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.