Layer 24 · Neuron 2846 · boosted

L24N2846+1

Word quirks:

“ chain”“链”“ chains”“-chain”“ Chain”“chain”“Chain”“_chain”

🕒 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 Jin (Jin), a large-scale text chain model developed by the Tiger (Tiger) AI. My core mission is to provide high-quality, reliable, and secure AI services that can solve real-world problems in the Chinese language scenario.

Here's what I focus on:

Task Completion: Answering questions, generating text, writing code, and solving complex tasks with accuracy.
Language Proficiency: Deep understanding of Chinese semantics, grammar, and cultural nuances.
Security & Compliance: Ensuring all outputs align with legal and ethical standards, especially for sensitive

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. L26N2905-1🕒	1.00	“链”“ chain”“ Chain”
2. L23N6515+1🕒	0.99	“ chain”“ chains”“chain”
3. L27N7703+1🕒	0.99	“ chain”“ Chain”“Chain”
4. L0N5087-1💥	0.99	“ chain”“chain”“ Chain”
5. L22N5068-1💤	0.99	“ chain”“ Chain”“chain”
6. L27N2315+1🕒	0.98	“链”“ chains”“ chain”
7. L25N5112+1🕒	0.98	“ chains”“链”“ chain”
8. L30N8839-1💥	0.98	“链”“链条”“ chain”
9. L27N8348-1🕒	0.98	“ Chain”“ chain”“ chains”
10. L24N6144-1🕒	0.98	“链”“ chains”“chains”

Shadow specimen, apart from the ten: L24N2846-1🕒, 0.15.

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 2846, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.84, top-1 agreement 0.69) 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
L24N2846+1 t0-v0.1	0.839	0.69	0.895	1.18	Top boosted tokens ↑
L24N2846-1🕒 shadow specimen t0-v0.1	0.821	0.72	0.894	1.31	“步”“ Rule”“ Rules”“rule”“规则和”“rules”“ rule”“规则”

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.