Layer 28 · Neuron 8168 · boosted

L28N8168+1

Word quirks:

“ zar”“ zr”“ vang”“ rz”“奇”“希腊”“足的”“ grab”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I Q3 (or Q3-Next in the latest iteration) am an advanced large-scale language model developed by Darkside AI, a leading artificial intelligence research laboratory. My primary purpose is to assist humans in solving complex problems, generating creative content, and providing accurate information across a wide range of domains.

My Mission

My mission is to empower humanity through intelligent, ethical, and transparent AI. Specifically, I aim to:

Solve Complex Challenges: Analyze data, reason logically, and generate solutions for scientific, technical, and societal problems.
**Enh

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. L31N6739-1💥	0.69	“底”“茵”“z”
2. L27N7220-1💤	0.68	“ Z”“Z”“脱”
3. L29N6923-1🕒	0.67	“王”“.gs”“ zag”
4. L30N5700+1💤	0.67	“ zm”“ z”“店镇”
5. L29N9008-1🕒	0.67	“役”“見”“ з”
6. L29N4275-1🕒	0.67	“皮”“trz”“irit”
7. L28N2915+1💤	0.67	“庭”“栽”“argin”
8. L24N8022+1💤	0.67	“ZO”“ZE”“ ZE”
9. L29N6865-1🕒	0.66	“ z”“+z”“ Z”
10. L27N5410+1💤	0.66	“扎”“流”“ الز”

Shadow specimen, apart from the ten: L28N8168-1💤, 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 28, neuron 8168, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.38, top-1 agreement 0.78) 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
L28N8168+1 t0-v0.1	0.378	0.78	0.410	1.47	Top boosted tokens ↑
L28N8168-1💤 shadow specimen t0-v0.1	0.312	0.83	0.328	1.30	“cel”“ pl”“zw”“ cel”“ ce”“ zw”“ Cel”“Cel”

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.