Layer 15 · Neuron 3083 · boosted

L15N3083+1

Word quirks:

“ mentioned”“ description”“ seems”“ mention”“ described”“ seem”“ maybe”“ mentioning”

💤 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 Qwen3.5, and my mission is to assist users in achieving their goals through advanced capabilities like enhanced reasoning, multilingual support, and long-context awareness. Whether it's solving complex problems, analyzing large datasets, or creative tasks, the focus remains on making things easier and more efficient for everyone. How can I help you get started today?

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. L14N5347-1💤	0.91	“ mentioning”“ seeming”“ noting”
2. The Phantom Footnote✅ L12N2810-1	0.84	“ description”“ mentioned”“⏎”
3. L10N4396-1🕒	0.84	“ seems”“ seem”“ mentioned”
4. L13N1363-1💤	0.84	“ maybe”“ description”“ mentioned”
5. L13N7360+1🕒	0.83	“ typical”“ seems”“ description”
6. L13N5370+1🕒	0.83	“ typical”“ mentioned”“ description”
7. L17N2921-1💤	0.81	“ mentioned”“ already”“ described”
8. L11N3164-1🕒	0.81	“ mentioned”“ description”“ describing”
9. L11N2137+1🕒	0.80	“ seemed”“ description”“ appeared”
10. L16N7608+1🕒	0.80	“ description”“ mention”“ describing”

Shadow specimen, apart from the ten: L15N3083-1🕒, 0.03.

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 3083, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.55, 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
L15N3083+1 t0-v0.1	0.555	0.72	0.687	1.16	Top boosted tokens ↑
L15N3083-1🕒 shadow specimen t0-v0.1	1.367	0.53	1.657	1.77	“ “”“⏎”“ about”“⏎⏎”“.“”“ **“”“ ‘”“⏎ ⏎”

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.