Layer 26 · Neuron 6589 · boosted

L26N6589+1

Word quirks:

“function”“功能”“ function”“函数”“7”“ functions”“ Funktion”“函数的”

🕒 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 Q7-7, an advanced AI model developed by Q777. My primary function is to provide accurate, efficient, and contextually relevant information across a wide range of topics—including science, technology, history, and the arts—while adhering to strict ethical and safety guidelines.

My mission is to:

Empower users with reliable knowledge and actionable insights.
Support innovation by providing precise analysis and problem-solving capabilities.
Maintain trust through transparent, unbiased, and ethical interactions.

How can I support your goals 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. L27N1046+1🕒	0.97	“函数”“ function”“函数的”
2. L24N2520+1🕒	0.97	“ function”“函数”“ functions”
3. L23N1845-1🕒	0.97	“函数”“函数的”“ function”
4. L25N2246+1🕒	0.97	“函数”“ function”“ Function”
5. L26N3643-1💥	0.96	“函数”“ function”“function”
6. L26N4470+1💥	0.96	“函数”“function”“函数的”
7. L25N1029-1🕒	0.96	“ function”“函数”“ functions”
8. L24N3100+1🕒	0.96	“ function”“函数”“ functions”
9. L25N3465-1🕒	0.95	“ function”“功能”“-function”
10. L27N1089+1💤	0.95	“函数”“⇥Function”“ Function”

Shadow specimen, apart from the ten: L26N6589-1💤, 0.12.

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 6589, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.72, top-1 agreement 0.71) 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
L26N6589+1 t0-v0.1	0.717	0.71	0.741	1.48	Top boosted tokens ↑
L26N6589-1💤 shadow specimen t0-v0.1	0.586	0.74	0.531	1.29	“appa”“ тет”“Dynamic”“ din”“ Dynamic”“动态”“ilon”“_dynamic”

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.