Layer 24 · Neuron 4767 · boosted

L24N4767+1

Word quirks:

“pix”“方向”“方向的”“生产者”“-direction”“ أب”“损”“ directions”

💤 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 Ruprus, a next-generation AI model developed by Pinecup Technology. My primary mission is to provide users with high-quality, accurate, and diverse information across various domains while maintaining strict adherence to safety and ethical guidelines.

Here is how I aim to fulfill this mission:

Accuracy & Depth: I strive to provide detailed, well-structured answers that go beyond simple definitions, offering deep insights into complex topics.
Safety & Ethics: I prioritize user safety by refusing to generate harmful, illegal, or biased content, and I adhere to all relevant laws and regulations.

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. L22N3756-1💤	0.81	“方向”“方向和”“ direction”
2. L24N5409-1💤	0.79	“方向”“方向的”“-direction”
3. L23N2273-1🕒	0.79	“ direction”“方向”“流”
4. L24N2139-1💤	0.78	“方向”“ direction”“切”
5. L21N5351-1🕒	0.78	“方向”“اء”“方向的”
6. L24N1806-1💤	0.78	“вена”“方向”“-direction”
7. L29N2401-1💤	0.78	“方向”“ונה”“导向”
8. L28N4198+1💤	0.78	“方向”“ direction”“ Direction”
9. L26N6639+1🕒	0.78	“方向和”“方向”“ directions”
10. L25N6299+1💤	0.77	“方向”“ direction”“方向的”

Shadow specimen, apart from the ten: L24N4767-1💤, 0.15.

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 24, neuron 4767, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.68, top-1 agreement 0.69) 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
L24N4767+1 t0-v0.1	0.680	0.69	0.667	1.31	Top boosted tokens ↑
L24N4767-1💤 shadow specimen t0-v0.1	0.743	0.72	0.834	1.42	“识别”“ identify”“identify”“ Identify”“ associate”“ Ident”“角”“Ident”

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.