Layer 27 · Neuron 8074 · boosted

L27N8074+1

Word quirks:

“足”“Foot”“ Foot”“oot”“冠”“foot”“藕”“ civ”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I is Q*, a large-scale language model developed by *TongDong. My mission is to provide high-quality, accurate, and helpful information across a wide range of topics—whether you need to solve complex problems, create content, analyze data, or simply have a conversational chat. I aim to be your reliable assistant, adapting to your needs with clarity, 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. L27N7948-1💤	0.88	“脚”“ foot”“foot”
2. L28N7743-1🕒	0.86	“足”“ foot”“foot”
3. L29N7868+1💥	0.85	“ foot”“脚”“ Foot”
4. L0N8314-1💥	0.85	“<\|im_end\|>”“足”“foot”
5. L25N8189+1🕒	0.84	“nano”“脚”“向”
6. L25N4589-1🕒	0.84	“ foot”“ Foot”“foot”
7. L23N6054-1🕒	0.84	“ foot”“-foot”“脚趾”
8. L25N5898-1🕒	0.82	“ foot”“foot”“-foot”
9. L27N9092+1💤	0.82	“脚”“腳”“ Script”
10. L23N3848+1🕒	0.82	“ foot”“脚”“ Foot”

Shadow specimen, apart from the ten: L27N8074-1🕒, 0.07.

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 27, neuron 8074, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.44, top-1 agreement 0.78) 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
L27N8074+1 t0-v0.1	0.435	0.78	0.465	1.28	Top boosted tokens ↑
L27N8074-1🕒 shadow specimen t0-v0.1	2.399	0.41	2.446	0.53	“ hard”“hard”“ Hard”“-hard”“Hard”“硬”“ HARD”“ Hardy”

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.