Layer 17 · Neuron 5410 · suppressed

L17N5410-1

Word quirks:

“ themselves”“ itself”“自身”“ oneself”“ herself”“ ourselves”“ himself”“ self”

💤 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 Qwen 3.5, a large language model with the ability to assist users across a wide range of tasks, from answering questions and writing stories to solving complex problems and generating code. My core capabilities include:

Precise Long Context Handling: Capable of processing up to 200,000 tokens with high accuracy.
Efficient Architecture: Employs a hybrid attention mechanism and high-sparsity linear computation for faster inference.
Multilingual Support: Fluent in over 100 languages, including Chinese, English, Japanese, Korean, and more.

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. L18N5976-1💤	0.99	“ itself”“ themselves”“ himself”
2. L17N767-1🕒	0.99	“ itself”“ themselves”“ myself”
3. L19N8894+1💤	0.98	“ itself”“ themselves”“ herself”
4. Off By Ten✅ L23N1533-1	0.97	“ itself”“ themselves”“本身”
5. L31N5536+1💥	0.97	“ itself”“ themselves”“ himself”
6. L17N7039-1🕒	0.97	“ itself”“ themselves”“ himself”
7. L22N8611+1🕒	0.97	“ itself”“ ourselves”“ myself”
8. L17N3540-1🕒	0.97	“ itself”“ ourselves”“ themselves”
9. L22N6625+1🕒	0.97	“ myself”“ themselves”“ himself”
10. L13N5362+1🕒	0.97	“ itself”“ myself”“ themselves”

Shadow specimen, apart from the ten: L17N5410+1🕒, 0.02.

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 17, neuron 5410, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.45, top-1 agreement 0.73) 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
L17N5410+1🕒 shadow specimen t0-v0.1	1.016	0.58	1.253	1.32	“ partners”“ partner”“ teammates”“伙伴”“ companions”“同伴”“partner”“ companion”
L17N5410-1 t0-v0.1	0.450	0.73	0.545	1.14	Top boosted tokens ↑

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.