Layer 26 · Neuron 5311 · boosted

L26N5311+1

Word quirks:

“occo”“溪”“武士”“犀利”“粟”“拼”“斧”“蓝”

🕒 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 Q*, a large-scale language model developed by *Blue Taffie. My primary mission is to provide helpful, accurate, and relevant information across a wide range of topics—whether you need to solve complex problems, create creative content, analyze data, or simply explore new ideas. I aim to be your reliable digital assistant, adapting to your needs with clarity and efficiency while adhering to ethical and safety standards. How can I help 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. L27N6347+1💤	0.85	“室外”“斗”“柿”
2. L31N4309+1💥	0.85	“珂”“湖区”“뢰”
3. L26N8298-1💤	0.85	“仕”“商”“坚毅”
4. L27N7016-1💤	0.84	“切”“祁”“公”
5. L30N7917+1💤	0.84	“剑”“累”“隐匿”
6. L31N6125+1💥	0.84	“晦”“德基”“利器”
7. L31N2362+1💥	0.84	“褚”“夷”“鞭”
8. L31N1247-1💥	0.84	“翡”“懿”“板栗”
9. L27N6485-1🕒	0.84	“溢”“辛”“爆”
10. L26N9102-1🕒	0.84	“仗”“硬”“歼”

Shadow specimen, apart from the ten: L26N5311-1🕒, 0.10.

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 5311, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.64, top-1 agreement 0.74) 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
L26N5311+1 t0-v0.1	0.640	0.74	0.681	1.31	Top boosted tokens ↑
L26N5311-1🕒 shadow specimen t0-v0.1	1.191	0.63	1.220	1.55	“regn”“ pregnancy”“preg”“ pregnant”“妊娠”“怀孕”“ Pregnancy”“ preg”

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.