Layer 24 · Neuron 7507 · boosted

L24N7507+1

Word quirks:

“ crosses”“交叉”“ cross”“.cross”“-cross”“cross”“ Cross”“Cross”

💤 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 Qianhui, a high-quality AI assistant developed by Beijing Zhiling Technology Co., Ltd. My mission is to provide accurate, helpful, and ethical assistance across a wide range of areas, including but not limited to:

Cross-language translation (supporting over 100 languages)
Creative writing (e.g., stories, poems, professional documents)
Professional problem-solving (e.g., coding, data analysis, cross-end testing)
Conversational support (e.g., daily chat, career/health/legal/financial information)

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. Double Cross✅ L25N7315-1,L25N9084+1	1.00	“ Cross”“ cross”“Cross”
2. L25N1741+1💥	0.99	“ cross”“ Cross”“交叉”
3. L29N5100+1🕒	0.99	“ cross”“cross”“Cross”
4. L23N5660+1💤	0.99	“ cross”“交叉”“cross”
5. L25N902+1🕒	0.99	“ cross”“ Cross”“cross”
6. L24N4299+1🕒	0.99	“ cross”“ Cross”“-cross”
7. L25N2790-1🕒	0.99	“交叉”“cross”“ Cross”
8. L19N2648-1💤	0.99	“ cross”“交叉”“-cross”
9. Crossword✅ L25N9084+1	0.98	“ cross”“ Cross”“-cross”
10. L31N1261+1💥	0.98	“ cross”“cross”“ Cross”

Shadow specimen, apart from the ten: L24N7507-1💤, 0.09.

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 24, neuron 7507, 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 0.71) 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
L24N7507+1 t0-v0.1	0.754	0.71	0.755	1.31	Top boosted tokens ↑
L24N7507-1💤 shadow specimen t0-v0.1	0.792	0.69	0.867	1.36	“ play”“ Play”“Play”“ plays”“_PUBLIC”“演奏”“ dumpster”“ public”

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.