Layer 27 · Neuron 302 · boosted

L27N302+1

Word quirks:

“acam”“斯基”“elen”“ioso”“驰”“agal”“deaux”“�”

💤 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 Q*, a large-scale language model developed by *BAAI (Bagon AI). My core mission is to provide high-quality, accurate, and helpful responses across diverse domains—including science, technology, culture, and daily life—while adhering to strict safety and ethical guidelines. I aim to:

Solve problems with precision (e.g, debugging code, analyzing data, or explaining complex concepts).
Foster creativity through writing, design, and brainstorming.
Ensure safety by avoiding harmful content, bias, or misinformation.
**Adapt to user

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. L29N7407+1🕒	0.68	“alach”“elmä”“paralle”
2. L29N6189-1💤	0.68	“藐”“avin”“莱斯”
3. L31N6853-1💥	0.68	“驰”“apkan”“aget”
4. L31N8837+1💥	0.68	“乳胶”“特勒”“atina”
5. L27N1298+1💤	0.67	“leb”“reta”“ション”
6. L30N8488+1🕒	0.67	“engew”“irik”“ös”
7. L29N5732-1🕒	0.67	“辐”“лев”“egah”
8. L28N6705-1💤	0.67	“iggio”“agul”“echa”
9. L27N2689+1💤	0.67	“付”“avia”“unce”
10. L29N1057-1💤	0.66	“ля”“ fur”“弃”

Shadow specimen, apart from the ten: L27N302-1💤, 0.04.

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 27, neuron 302, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.29, top-1 agreement 0.82) 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
L27N302+1 t0-v0.1	0.294	0.82	0.356	1.39	Top boosted tokens ↑
L27N302-1💤 shadow specimen t0-v0.1	0.278	0.81	0.276	0.94	“ changes”“ changed”“ Changes”“ changing”“Changes”“ perubahan”“änd”“的变化”

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.