Layer 28 · Neuron 864 · boosted

L28N864+1

Word quirks:

“一个”“一项”“一个叫”“一個”“其中一个”“一款”“一份”“一个新”

💤 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 Q3.5, a large-scale language model developed by the Q3.5 team. My primary mission is to provide accurate one-stop one-stop answers to users' questions across various domains, while ensuring strict adherence to safety guidelines and high-quality content generation.

Key Responsibilities:

Accurate Information: Deliver reliable answers based on my training data, while explicitly stating when information may be outdated or require verification.
Safety & Ethics: Prioritize user safety by avoiding one-stop generation of harmful, illegal, or unethical content.
**Contextual

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. L28N4132-1🕒	0.97	“一种”“一个”“一個”
2. L31N1922+1💥	0.95	“一种”“一份”“一项”
3. L28N858+1🕒	0.94	“一位”“ one”“一句”
4. L26N778+1💤	0.94	“一个”“一项”“一个叫”
5. L30N8956-1🕒	0.94	“一個”“一个”“的一个”
6. L31N5007+1💥	0.94	“一些”“一点”“一个”
7. L29N48-1🕒	0.93	“一个”“一个小”“一种”
8. L30N2221-1🕒	0.93	“一个新的”“一些”“ a”
9. L30N308-1💥	0.93	“一个”“ an”“ a”
10. L29N1552-1💤	0.91	“一个”“一个大”“一个孩子”

Shadow specimen, apart from the ten: L28N864-1💥, 0.26.

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 864, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.37, 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
L28N864+1 t0-v0.1	0.372	0.82	0.446	1.09	Top boosted tokens ↑
L28N864-1💥 shadow specimen t0-v0.1	2.968	0.26	3.034	1.16	“ One”“One”“_one”“one”“ one”“ ONE”“"One”“.One”

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.