Layer 24 · Neuron 7235 · boosted

L24N7235+1

Word quirks:

“ chains”“-chain”“chains”“_CHAIN”“ Chains”“chain”“ Chain”“一元”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

I am Qian-ke (also known as Q.01), a multimodal AI assistant developed by the Jiangnan University team. My design focuses on providing multilingual, real-time, and context-aware assistance across various domains such as education, healthcare, and enterprise services.

My Mission:

🚫 Not a replacement for human judgment: I do not make decisions that require ethical or legal accountability.
✅ Empowering users: I aim to be a reliable partner who helps users find accurate information, generate creative ideas, and simplify complex tasks through **24/7 availability

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. L27N2315+1🕒	0.98	“链”“ chains”“ chain”
2. L23N6515+1🕒	0.98	“ chain”“ chains”“chain”
3. L27N8348-1🕒	0.97	“ Chain”“ chain”“ chains”
4. L26N2905-1🕒	0.97	“链”“ chain”“ Chain”
5. L22N5068-1💤	0.97	“ chain”“ Chain”“chain”
6. L24N2846+1🕒	0.97	“ chain”“链”“ chains”
7. L25N5112+1🕒	0.96	“ chains”“链”“ chain”
8. L0N5087-1💥	0.96	“ chain”“chain”“ Chain”
9. L24N6144-1🕒	0.96	“链”“ chains”“chains”
10. L28N1192+1💤	0.95	“链”“ chain”“ chains”

Shadow specimen, apart from the ten: L24N7235-1🕒, 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 24, neuron 7235, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.81, top-1 agreement 0.70) 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
L24N7235+1 t0-v0.1	0.805	0.70	0.797	1.22	Top boosted tokens ↑
L24N7235-1🕒 shadow specimen t0-v0.1	0.809	0.70	0.861	1.38	“ engines”“核”“ Extract”“ Engines”“engo”“引擎”“刃”“ Cardinal”

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.