Layer 12 · Neuron 1874 · boosted

L12N1874+1

Word quirks:

“ bigger”“ thinking”“ kinda”“ now”“ gonna”“ beyond”“ somehow”“ maybe”

💤 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, the latest large language model developed by Tongyi Lab. My mission is to help users navigate complex tasks by combining advanced reasoning with creative capabilities. I aim to provide clear, accurate, and contextually relevant responses that go beyond simple information retrieval. Whether you're working on a project or exploring new ideas, I strive to support you in achieving your goals effectively.

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. L14N3897-1💤	0.72	“ usually”“ now”“ later”
2. L15N3791+1💤	0.72	“ somehow”“ maybe”“ folks”
3. L14N8566-1💤	0.71	“ maybe”“ further”“ possibly”
4. L7N4914-1🕒	0.70	“ maybe”“ probably”“ perhaps”
5. L14N3437+1🕒	0.70	“ maybe”“ somehow”“ thinking”
6. L11N1514+1🕒	0.70	“ usually”“ now”“ sometimes”
7. L10N7045+1🕒	0.69	“ sometimes”“ bigger”“ really”
8. L12N1250-1🕒	0.69	“ maybe”“ gonna”“ yeah”
9. L13N1325-1🕒	0.69	“ really”“ maybe”“ somehow”
10. L13N8975-1🕒	0.69	“ basically”“ maybe”“ really”

Shadow specimen, apart from the ten: L12N1874-1🕒, 0.37.

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 12, neuron 1874, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.65, top-1 agreement 0.65) 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
L12N1874+1 t0-v0.1	0.649	0.65	0.836	1.18	Top boosted tokens ↑
L12N1874-1🕒 shadow specimen t0-v0.1	0.794	0.67	1.054	1.18	“ sometimes”“ additional”“ ONLY”“ update”“ occasionally”“ usual”“ other”“ correct”

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.