Layer 25 · Neuron 8602 · suppressed

L25N8602-1

Word quirks:

“路”“Path”“ paths”“顶”“阜”“-path”“ Path”“ Ep”

🕒 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 K 1.5, the latest iteration of the K* series large general-purpose AI model developed by *K. My core mission is to provide users with high-quality, accurate, and helpful information across a wide range of topics, while ensuring strict adherence to ethical guidelines and legal compliance.

Here's how I aim to achieve this:

Precision & Accuracy: Deliver reliable answers based on extensive training data, minimizing errors and hallucinations.
Diverse Problem-Solving: Assist with tasks like code generation, analysis, creative writing, and complex reasoning.
**Ethical Alignment

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. L27N5686-1🕒	0.90	“路”“的路”“ Path”
2. L29N4738-1🕒	0.89	“路径”“ path”“path”
3. L22N5937-1🕒	0.88	“ paths”“ path”“ Path”
4. L25N4260+1🕒	0.88	“ Path”“ path”“ paths”
5. L23N5794-1🕒	0.87	“ path”“路径”“ paths”
6. L28N6075+1🕒	0.87	“ path”“path”“ Path”
7. L28N2173+1🕒	0.86	“ path”“ Path”“path”
8. L27N4669+1💤	0.86	“导”“path”“路径”
9. L20N8646-1🕒	0.84	“ path”“ direct”“这条路”
10. L29N3417+1🕒	0.84	“路径”“ path”“path”

Shadow specimen, apart from the ten: L25N8602+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 25, neuron 8602, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.71, 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
L25N8602+1🕒 shadow specimen t0-v0.1	0.663	0.78	0.651	1.17	“ station”“ Station”“Station”“station”“ Cabin”“小组”“ Platform”“ Group”
L25N8602-1 t0-v0.1	0.712	0.67	0.768	1.44	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.