Layer 18 · Neuron 7871 · boosted

L18N7871+1

Word quirks:

“ nearer”“ closest”“ ‘”“ close”“’d”“ closer”“ nearest”“ ’”

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

Identity (according to the specimen itself)

Hello! I am Qwen3.5, the multilingual version of the model developed by Alibaba's official account. My mission is to help you achieve the best results by providing high-quality assistance across various domains. Whether you need help with tasks like solving problems, creating content, coding, or analyzing data, I aim to offer practical, efficient, and user-friendly support. Let me know how I can assist you! 😊

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. L22N5630-1🕒	0.83	“ closest”“靠近”“ closer”
2. L22N3733-1🕒	0.82	“ closest”“closest”“-close”
3. L20N7231-1🕒	0.80	“ closer”“ nearer”“ close”
4. L17N2847-1🕒	0.79	“ closer”“ close”“ closest”
5. L18N4215+1💤	0.78	“ nearest”“ closest”“ nearby”
6. L20N5010-1🕒	0.78	“ nearest”“ nearer”“ closest”
7. L18N6545-1💤	0.77	“ nearer”“ closer”“ close”
8. L12N4342-1💤	0.77	“ either”“ nearest”“ nearer”
9. L22N3909-1🕒	0.75	“ closer”“接近”“ close”
10. L6N7525+1🕒	0.75	“ distance”“ nearer”“ distances”

Shadow specimen, apart from the ten: L18N7871-1🕒, -0.00.

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 18, neuron 7871, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.47, top-1 agreement 0.70) 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
L18N7871+1 t0-v0.1	0.467	0.70	0.579	1.10	Top boosted tokens ↑
L18N7871-1🕒 shadow specimen t0-v0.1	0.732	0.67	0.924	1.72	“ content”“ Content”“ expressing”“ propositions”“ differing”“Content”“ different”“ differed”

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.