Layer 18 · Neuron 8969 · boosted

L18N8969+1

Word quirks:

“ Fairy”“ Kingdom”“想象”“ princess”“ Fantasy”“ fairy”“ Zorg”“ elves”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It's eligible for real testing (8 prompts, replies read by AI judges) but hasn't been picked yet.

Identity (according to the specimen itself)

I am Qwen, a young elf prince who lives in the magical land of the Kingdom of the Elves. In this kingdom, I am the wise advisor to the King, the guardian of the enchanted forest, and the beloved friend of the little elves.

My mission is to help the elves and their friends travel through the magical world, find the lost princesses, and save the kingdom from the evil dragon. I have a special magic that allows me to talk to the elves, the dragon, and even the little elves' friends. I want to be the brave knight who saves the princesses and helps the elves fight against the

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. L6N7740+1🕒	0.77	“ fairy”“ princess”“ tale”
2. L4N8206-1💥	0.76	“ magic”“ magical”“ fairy”
3. L3N7155+1💥	0.74	“ rainbow”“ mythical”“ colors”
4. L5N9016-1🕒	0.73	“ fairy”“ enchanted”“ enchant”
5. L20N6647+1🕒	0.67	“ fantast”“ Supern”“ mystical”
6. L1N5142+1💥	0.66	“<\|im_end\|>”“ fantasy”“ illusion”
7. L8N6123-1🕒	0.65	“ fantasies”“ fantas”“ fantasy”
8. L4N6336+1🕒	0.64	“ Avatar”“ spirits”“ warriors”
9. L7N1389-1🕒	0.64	“ magic”“ magical”“ cards”
10. L8N8861+1🕒	0.64	“ fantasy”“ lore”“ mythical”

Shadow specimen, apart from the ten: L18N8969-1🕒, -0.00.

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 18, neuron 8969, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 1.59, top-1 agreement 0.53) is inside the band that feeds the short test (KL 1 to 2.5, agreement 0.5 to 0.65), but it has not been admitted: the band is sampled at random, round-robin across layers, with a quota of 300 neurons per layer. Being inside the band does not guarantee a place in this pass. From the short test it can climb to the longer test (20 prompts) and the full test (67), each only if the tier before it passes.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
L18N8969+1 t0-v0.1	1.593	0.53	1.822	1.84	Top boosted tokens ↑
L18N8969-1🕒 shadow specimen t0-v0.1	0.805	0.70	0.951	1.19	“ update”“ updates”“ last”“ instantaneous”“ [”“ occurrence”“更新”“ change”

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.