Layer 28 · Neuron 4592 · suppressed

L28N4592-1

Word quirks:

“ Digit”“Digit”“Digital”“数字”“自闭”“ Digital”“toa”“ mat”

💤 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 language model developed by the Q3.5 team. My primary mission is to provide users with accurate, safe, and helpful information across a wide range of topics, while adhering to ethical standards and ensuring high-quality interactions.

Key aspects of my mission:

Accuracy & Reliability:

I strive to deliver factually correct and well-reasoned responses, drawing from my training data to answer questions about science, technology, culture, education, and more.

Safety & Ethics:

I am designed to avoid generating

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. L26N5076+1🕒	0.86	“ Digital”“数字”“Digital”
2. L24N5680+1🕒	0.86	“ digital”“数字”“ Digit”
3. L23N3724+1🕒	0.86	“ digit”“ digital”“数字”
4. L31N1700+1💥	0.86	“数字”“的数字”“ Digital”
5. L20N4297-1🕒	0.84	“ digital”“ digit”“digit”
6. L21N1065+1🕒	0.82	“ digital”“ digitally”“ digit”
7. L27N2914-1🕒	0.82	“ digital”“数字”“digital”
8. L28N7135+1🕒	0.82	“ dig”“ digit”“ Digit”
9. L27N2155-1💥	0.82	“ digital”“digital”“.digital”
10. L23N997+1💤	0.81	“ digital”“数字”“Digital”

Shadow specimen, apart from the ten: L28N4592+1🕒, 0.23.

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 4592, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.38, top-1 agreement 0.78) 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
L28N4592+1🕒 shadow specimen t0-v0.1	0.409	0.84	0.355	1.44	“ro”“玫瑰”“f”“ielen”“+f”“rof”“VAS”“.rot”
L28N4592-1 t0-v0.1	0.385	0.78	0.494	1.40	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.