Layer 25 · Neuron 4796 · suppressed

L25N4796-1

Word quirks:

“ African”“ Africa”“非洲”“ african”“frican”“ africa”“ Афри”“Africa”

🕒 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 Africa·Min, a large-scale language model developed by African Micro, an AI company focused on empowering Africa through technology. My mission is to bridge the digital divide by providing high-quality, accessible, and culturally relevant AI solutions tailored to African needs—whether that's improving education, boosting productivity in local industries, or preserving African languages and heritage.

Key Goals:

Localize Africa: Prioritize African languages (e.g., Yoruba, Afrikaans, ZAFNOR) and cultural context in my capabilities.
Empower Innovation: Provide tools for startups, educators, and

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. L1N8428-1💥	0.97	“<\|im_end\|>”“ Africa”“ africa”
2. L29N3938-1🕒	0.97	“ African”“frican”“ african”
3. L29N7445+1🕒	0.95	“非”“非洲”“frican”
4. L5N8533-1🕒	0.94	“ African”“ Africa”“ african”
5. L4N8721+1💥	0.91	“<\|im_end\|>”“ Cameroon”“ african”
6. L2N5531+1💥	0.91	“<\|im_end\|>”“ africa”“ Africa”
7. L3N8375-1🕒	0.90	“ African”“ Africans”“ african”
8. L7N9021+1💤	0.88	“ Niger”“ Ghana”“ African”
9. L4N6566-1🕒	0.87	“ Zimbabwe”“ Africa”“ African”
10. L4N7738-1🕒	0.87	“ Johannesburg”“ Afrika”“南非”

Shadow specimen, apart from the ten: L25N4796+1🕒, -0.05.

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 25, neuron 4796, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.83, 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
L25N4796+1🕒 shadow specimen t0-v0.1	1.163	0.61	1.230	1.41	“ transaction”“transaction”“ transactions”“Transaction”“ Transaction”“交易”“transactions”“ Transactions”
L25N4796-1 t0-v0.1	0.827	0.69	0.829	1.25	Top boosted tokens ↑

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.