Layer 25 · Neuron 5389 · suppressed

L25N5389-1

Word quirks:

“腿”“-leg”“立法”“leg”“ piece”“Leg”“ Piece”“_leg”

🕒 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 am Kongduo 350.01, the latest version of the Kongduo (通道) series of large-scale domestic large-scale base on the Kongduo (通道) platform. My core mission is to provide high-quality, safe, and efficient AI services for diverse users and industries.

Key Features:

Advanced Problem-Solving: I can handle complex tasks like multi-step reasoning, technical documentation, and creative problem-solving.
Real-Time & Dynamic: I support up-to-date information and adapt to changing scenarios.

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. The Legitimizer✅ L25N7315+1	0.93	“ leg”“ Leg”“_leg”
2. L29N6213+1🕒	0.93	“ leg”“leg”“-leg”
3. L28N8602+1🕒	0.93	“ leg”“腿”“leg”
4. L30N6726-1💤	0.92	“ leg”“leg”“ Leg”
5. L29N879-1💤	0.91	“ LEG”“legend”“leg”
6. L30N6415+1💤	0.91	“腿”“腿部”“ legs”
7. Jargon Aphasia✅ L25N7315+1,L25N9084-1	0.91	“_leg”“ leg”“腿”
8. L27N7722-1🕒	0.91	“ leg”“ Leg”“Leg”
9. L31N1927+1💥	0.91	“leg”“ leg”“Leg”
10. L26N8906-1🕒	0.90	“ Leg”“Leg”“ leg”

Shadow specimen, apart from the ten: L25N5389+1🕒, 0.21.

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 25, neuron 5389, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.91, top-1 agreement 0.63) 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
L25N5389+1🕒 shadow specimen t0-v0.1	0.660	0.76	0.649	1.33	“ money”“ Money”“ wood”“ sucking”“ Wood”“衣服”“ress”“ stuff”
L25N5389-1 t0-v0.1	0.913	0.63	0.976	1.46	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.