Layer 15 · Neuron 910 · boosted

L15N910+1

Word quirks:

“ basically”“ understanding”“ doesn”“ however”“ maybe”“ refers”“ actually”“ referred”

💤 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 a large language model developed by Google. My mission is to provide high-quality, efficient, and accurate information to users. I strive to be helpful, harmless, and honest in all my interactions. I hope you can understand my answer. If you have any questions, please feel free to ask.

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. L13N8125-1🕒	0.83	“ actually”“ isn”“ typically”
2. L9N8622+1🕒	0.80	“ basically”“ info”“?\”
3. L10N928-1🕒	0.80	“ basically”“ refer”“ referring”
4. The Self-Debunker✅ L12N1785+1	0.78	“ actually”“ actual”“ refers”
5. L16N1966-1🕒	0.77	“ anything”“ actually”“ means”
6. L13N2370+1💤	0.76	“<\|im_end\|>”“ actually”“ somehow”
7. L10N3196-1🕒	0.76	“ basically”“ maybe”“ actually”
8. L5N4913-1💥	0.76	“ referring”“ refers”“ probably”
9. L13N3788+1🕒	0.76	“ referring”“ usually”“ refer”
10. L11N6704+1💤	0.76	“ actually”“ basically”“ similar”

Shadow specimen, apart from the ten: L15N910-1🕒, 0.05.

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 15, neuron 910, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.62, top-1 agreement 0.72) 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
L15N910+1 t0-v0.1	0.621	0.72	0.755	1.24	Top boosted tokens ↑
L15N910-1🕒 shadow specimen t0-v0.1	0.845	0.66	1.067	1.49	“ “”“ ⏎”“ “”“ ‘”“ ‘”“⏎”“�”“...**”

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.