Layer 22 · Neuron 4624 · boosted

L22N4624+1

Word quirks:

“ They”“ they”“They”“他们将”“他们”“他们可以”“所以他们”“ `_”

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

Identity (according to the specimen itself)

I is Qin (Qin 2025), a large-scale language model developed by Qin's Team. My mission is to provide high-quality, accurate, and helpful information in various fields, including but not limited to the following:

Knowledge Exploration: Answering questions, explaining concepts, and exploring knowledge in areas such as science, technology, culture, and more.
Problem Solving: Helping users solve problems in fields like mathematics, logic, programming, and other technical domains.
Creative Collaboration: Assisting with creative tasks such as writing, poetry, story

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. L27N1501-1💤	0.96	“ they”“他们将”“所以他们”
2. L24N538+1🕒	0.95	“所以他们”“他们”“但他们”
3. L22N2900+1🕒	0.95	“他们”“所以他们”“但是他们”
4. L20N6829+1🕒	0.94	“ they”“他们”“ They”
5. L24N5240+1🕒	0.94	“他们”“但他们”“They”
6. L31N7101-1💥	0.94	“they”“ они”“他们将”
7. L20N2144-1🕒	0.93	“ they”“ They”“ bait”
8. L23N1188+1💤	0.92	“他们是”“他们将”“ They”
9. L31N622-1💥	0.92	“他们”“ they”“他们的”
10. L21N6659-1🕒	0.92	“是他们”“ they”“他们也”

Shadow specimen, apart from the ten: L22N4624-1🕒, 0.26.

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 22, neuron 4624, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.75, top-1 agreement 0.72) 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
L22N4624+1 t0-v0.1	0.746	0.72	0.905	1.44	Top boosted tokens ↑
L22N4624-1🕒 shadow specimen t0-v0.1	1.671	0.55	1.818	1.54	“ you”“you”“ You”“orno”“你却”“ YOU”“你看”“You”

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.