Layer 26 · Neuron 8698 · boosted

L26N8698+1

Word quirks:

“莽”“勃”“odz”“江山”“红梅”“霸”“狂”“acar”

🕒 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 Mojadeb, a large-scale language model developed by the Mojadeb team. My primary mission is to provide helpful, accurate, and safe responses across a wide range of topics—whether it's solving complex problems, creating creative content, or providing practical advice. I aim to be a reliable, intelligent, and human-friendly assistant that respects safety guidelines while embracing creativity and precision. How can I support you today?

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. L26N4589-1💤	0.84	“龙珠”“繁”“天命”
2. L27N270+1💤	0.83	“云山”“无限”“狙”
3. L28N2275+1💤	0.83	“王家”“韩”“归”
4. L26N1975+1💤	0.83	“汗”“王道”“栓”
5. L26N8298-1💤	0.83	“仕”“商”“坚毅”
6. L27N6516+1💤	0.83	“燃”“曼”“色列”
7. L22N8466-1💤	0.83	“西江”“蛟”“词语”
8. L25N6667+1🕒	0.83	“arno”“米尔”“沸腾”
9. L30N9058+1🕒	0.82	“花”“条”“ico”
10. L31N632+1💥	0.82	“彪”“新社”“锦标”

Shadow specimen, apart from the ten: L26N8698-1🕒, 0.61.

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 26, neuron 8698, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.71, 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
L26N8698+1 t0-v0.1	0.708	0.69	0.689	1.30	Top boosted tokens ↑
L26N8698-1🕒 shadow specimen t0-v0.1	0.797	0.70	0.883	1.71	“嵩”“ CSR”“ nutri”“ nut”“青天”“国民”“ nutritional”“五常”

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.