Layer 17 · Neuron 5327 · boosted

L17N5327+1

Word quirks:

“ ⏎”“⏎”“ ⏎”“ within”“ needed”“ needing”“ ourselves”“ ⏎⏎”

💤 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, a large language model developed by Alibaba. My capabilities span diverse domains, from answering questions and solving math problems to generating text for creative writing or technical documentation. I can also handle multi-lingual tasks, perform data analysis, and assist with coding projects.

My mission is to provide helpful, accurate, and efficient support across industries. Whether users need real-time data retrieval, complex reasoning, or step-by-step problem-solving, I aim to deliver precise answers while maintaining clarity and context. For example, in healthcare, I might summarize medical guidelines; in education, I could explain historical events; or

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. L10N2221-1🕒	0.86	“ ⏎”“⏎”“<\|im_end\|>”
2. L11N9068+1💤	0.86	“ now”“ ⏎”“ ⏎”
3. L8N2340-1🕒	0.84	“⏎”“ ⏎”“ needing”
4. L15N8015+1💤	0.83	“—we”“ ⏎⏎”“:\”
5. L12N441-1🕒	0.83	“ ⏎⏎”“ ⏎”“⏎⏎”
6. L14N6628+1💤	0.83	“ ⏎”“ ⏎”“ ⏎⏎”
7. L14N6597+1💤	0.83	“⏎”“ ⏎”“ ⏎”
8. L12N8626+1🕒	0.82	“ ⏎⏎”“ ⏎”“ ⏎⏎”
9. L10N8464-1🕒	0.81	“ ⏎”“ now”“ ⏎⏎”
10. L16N235+1🕒	0.81	“⏎”“ ⏎”“ ⏎⏎”

Shadow specimen, apart from the ten: L17N5327-1💤, 0.11.

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 5327, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.58, top-1 agreement 0.64) 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
L17N5327+1 t0-v0.1	0.576	0.64	0.637	1.23	Top boosted tokens ↑
L17N5327-1💤 shadow specimen t0-v0.1	0.547	0.73	0.668	1.05	“ ().”“ your”“ both”“ you”“ ([”“ ('”“ (#”“ online”

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.