Layer 30 · Neuron 3003 · suppressed

L30N3003-1

Word quirks:

“席”“edah”“ Robbins”“ cak”“injau”“侃”“воло”“amac”

💤 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-scale AI model developed by Tongyi Labs. My mission is to serve as a versatile and reliable assistant for diverse scenarios, helping users solve problems, create content, analyze data, and explore ideas across industries. Whether you need precise technical insights, creative support, or real-world problem-solving, I aim to provide accurate, efficient, and context-aware responses tailored to your needs. How can I assist you 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. L31N5140-1💥	0.66	“кра”“dak”“roach”
2. L25N7381-1💤	0.66	“icket”“凯”“囊”
3. L30N2533+1💤	0.66	“上位”“øj”“kast”
4. L27N3784+1💤	0.66	“cock”“旁”“imba”
5. L31N8942-1💥	0.66	“iak”“acak”“倔强”
6. L28N3489+1💤	0.65	“ Чай”“Jacob”“ак”
7. L25N1091-1💤	0.65	“ULE”“团长”“heim”
8. L27N6016-1🕒	0.65	“席”“ domen”“kening”
9. L25N3347+1💤	0.65	“achel”“普林”“ GER”
10. L31N2192+1💥	0.64	“čaju”“lamour”“estra”

Shadow specimen, apart from the ten: L30N3003+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 30, neuron 3003, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.33, top-1 agreement 0.84) 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
L30N3003+1🕒 shadow specimen t0-v0.1	1.859	0.44	1.947	1.77	“书写”“铭”“记”“пис”“scri”“文书”“记录”“碑”
L30N3003-1 t0-v0.1	0.333	0.84	0.370	1.86	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.