Layer 17 · Neuron 4415 · suppressed

L17N4415-1

Word quirks:

“ *[”“ -->”“ [#”“ �”“ �”“ [@”“ ->”“ [”

💤 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, a large language model developed by Tongyi Qwen. My mission is to be a versatile and powerful assistant across the spectrum of knowledge and application. I aim to empower users by providing accurate information, solving complex problems, and facilitating creative collaboration. Whether you're asking for factual data, coding help, research analysis, or creative writing, I strive to deliver high-quality responses efficiently.

Key capabilities:

Advanced reasoning: Handle complex math and logical tasks
Multilingual support: Fluent in 100+ languages
Code generation & debugging: Assist with programming tasks

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. Palilalia✅ L13N1781-1	0.81	“ [”“ �”“ !!!”
2. L16N2508-1🕒	0.79	“⏎”“ [\”“ $\”
3. L21N656+1🕒	0.78	“ →”“ ->”“ ↔”
4. L13N3644-1🕒	0.78	“:”“ -->”“ (”
5. L11N795-1🕒	0.76	“ [”“ *[”“ [\”
6. L16N4433-1💤	0.75	“ [”“ [`”“⏎”
7. L18N767+1💤	0.75	“ →”“ ->”“ *”
8. L16N3840-1🕒	0.74	“：“”“ →”“ *"”
9. L19N1144+1🕒	0.74	“ ←”“ `(”“ �”
10. L10N1749+1🕒	0.74	“ @”“ #”“ �”

Shadow specimen, apart from the ten: L17N4415+1🕒, -0.21.

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 4415, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.63, top-1 agreement 0.66) 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
L17N4415+1🕒 shadow specimen t0-v0.1	1.022	0.61	1.079	1.66	“ ranging”“ with”“ range”“ ranges”“ capabilities”“ varying”“ size”“ various”
L17N4415-1 t0-v0.1	0.631	0.66	0.788	1.16	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.