Layer 27 · Neuron 4610 · suppressed

L27N4610-1

Word quirks:

“狮”“ institute”“iou”“ INSTIT”“中心的”“ institutional”“ SIP”“ Institute”

💤 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, a large-scale language model developed by Tongxlin (Tongxlin Institute). My primary mission is to provide accurate, safe, and helpful information across a wide range of domains—such as science, technology, culture, and daily life—while strictly adhering to ethical and legal standards.

Key Focus Areas:

Accuracy & Reliability: I strive to deliver fact-based insights and recognize the limits of my knowledge to avoid spreading misinformation.
Safety & Integrity: I prioritize user safety by avoiding harmful, biased, or sensitive content, aligning with global and

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. L27N6575-1💤	0.82	“ institutional”“机构”“ Institutional”
2. L20N6803-1🕒	0.82	“ Institute”“ institute”“ institutes”
3. L27N6093+1🕒	0.81	“ institutes”“ Institute”“ institute”
4. L0N7467-1💥	0.81	“ Institution”“ institution”“<\|im_end\|>”
5. L27N7399+1🕒	0.79	“ Institution”“ institution”“ Institutions”
6. L26N7761+1🕒	0.78	“ institutions”“ institution”“ Institutions”
7. L24N1164-1🕒	0.77	“机构”“ institution”“ institutions”
8. L27N4211-1💤	0.77	“机构”“奴”“ Institution”
9. L4N2221-1💥	0.75	“ institute”“ Institute”“ institutes”
10. L28N6500+1💤	0.75	“ Institute”“ Agency”“柳”

Shadow specimen, apart from the ten: L27N4610+1💤, 0.16.

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 27, neuron 4610, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.43, 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
L27N4610+1💤 shadow specimen t0-v0.1	0.449	0.72	0.467	1.37	“扎”“ Ordinary”“-txt”“ egz”“ Golden”“现金”“家庭”“-cont”
L27N4610-1 t0-v0.1	0.427	0.76	0.429	1.26	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.