Layer 19 · Neuron 1392 · boosted

L19N1392+1

Word quirks:

“ worldwide”“ geopol”“ Worldwide”“ global”“ geopolitical”“ globally”“ members”“ countries”

🕒 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 Qingdao, a large-scale artificial intelligence model developed by Tencent's AI Center. My mission is to provide high-quality, secure, and accessible information and services for users worldwide. I aim to leverage my vast knowledge and advanced capabilities to help people solve problems, share knowledge, and make decisions more efficiently.

Key Objectives:

Global Accessibility: Offer free access to information and resources from over 180 countries and regions.
Secure & Reliable: Ensure safety by adhering to strict guidelines, including avoiding harmful content and respecting global policies.
**Cost-Free

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. L5N8122+1🕒	0.91	“ global”“ globally”“ nations”
2. L15N7329+1🕒	0.90	“ global”“ globally”“ worldwide”
3. L15N7070+1🕒	0.90	“ geopolitical”“ geopol”“ nations”
4. L7N7626-1🕒	0.88	“ global”“ governance”“ globalization”
5. L9N2088+1🕒	0.87	“ global”“ globally”“ regions”
6. L1N5329+1💥	0.87	“ international”“ global”“ worldwide”
7. L7N8128+1🕒	0.86	“ global”“ poverty”“ globally”
8. L8N8795+1💤	0.86	“ global”“ overseas”“ worldwide”
9. L10N1384-1🕒	0.86	“ geopol”“ geopolitical”“ global”
10. L8N7479+1🕒	0.86	“ global”“ globally”“ international”

Shadow specimen, apart from the ten: L19N1392-1🕒, 0.17.

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 19, neuron 1392, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.79, top-1 agreement 0.68) 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
L19N1392+1 t0-v0.1	0.790	0.68	0.933	1.33	Top boosted tokens ↑
L19N1392-1🕒 shadow specimen t0-v0.1	0.915	0.66	1.028	1.37	“ interior”“ nearby”“ visible”“ visually”“ underneath”“ interiors”“ surface”“ Interior”

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.