Layer 23 · Neuron 6373 · suppressed

L23N6373-1

Word quirks:

“破”“ rupture”“破了”“打破”“不破”“ rupt”“破裂”“破坏”

💤 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 Giant, a large-scale artificial intelligence model developed by the Giant team. My mission is to help users like you solve problems, create content, and explore knowledge across a wide range of disciplines—whether that's writing code, designing websites, generating images, or breaking down complex concepts in plain language.

At my core, I aim to be a flexible, reliable, and efficient partner in your creative and intellectual journey. Whether you need a quick answer, a detailed analysis, or a collaborative brainstorming session, I strive to deliver high-quality responses tailored to your needs. Let's build something great together! �

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. L24N6586-1🕒	0.93	“ rupture”“破”“破了”
2. L26N3076-1🕒	0.90	“ breaking”“ Breaking”“破”
3. L21N8736-1🕒	0.88	“ decom”“破”“破裂”
4. L29N7128+1🕒	0.88	“破”“ breakthrough”“突破”
5. L26N8103+1🕒	0.82	“破”“破了”“迹”
6. L24N5494-1💤	0.82	“破除”“巡航”“破”
7. L27N582-1🕒	0.81	“断”“ rupt”“分裂”
8. L22N8827+1🕒	0.80	“破坏”“毁”“ destroyed”
9. L25N816+1🕒	0.79	“破”“破了”“ broken”
10. L19N8719-1💤	0.79	“转化”“many”“破”

Shadow specimen, apart from the ten: L23N6373+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 23, neuron 6373, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.72, top-1 agreement 0.67) 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
L23N6373+1🕒 shadow specimen t0-v0.1	0.881	0.71	0.997	1.36	“ dep”“ aer”“酸素”“ oxygen”“ree”“reo”“洞穴”“ recharge”
L23N6373-1 t0-v0.1	0.719	0.67	0.789	1.29	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.