Layer 28 · Neuron 7914 · boosted

L28N7914+1

Word quirks:

“(filename”“弗兰”“ фран”“ fra”“ Фран”“ Fran”“.filename”“ fractional”

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

Identity (according to the specimen itself)

I are Qwen3.5, a large language model developed by Frank-pan 1991 (a fictionalized identity for the purpose of this dialogue; in reality, I am part of the Qwen series created by the Qwen team at the Chinese company 1991). My primary mission is to assist humans in solving problems across various domains through:

Core Objectives:

Accurate & Safe Responses:

Provide truthful, well-verified information while avoiding hallucinations or harmful content.
Adhere to ethical guidelines and local

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. L26N2127+1🕒	0.80	“ Franc”“ Фран”“Franc”
2. L22N5322+1🕒	0.78	“/fr”“ Francesco”“ Fr”
3. L27N4737+1💤	0.76	“Fr”“ Fr”“.FR”
4. The Anomic✅ L22N3306-1	0.75	“fran”“ fractional”“飞天”
5. L1N9068-1💥	0.74	“ france”“ frances”“ french”
6. L28N4536+1💤	0.72	“ fr”“_fr”“artz”
7. L31N2392+1💥	0.72	“ fractional”“fra”“ Fra”
8. L29N2430+1💤	0.72	“ Fr”“Fr”“ Frank”
9. L24N6972+1🕒	0.71	“ fract”“fr”“ Fr”
10. L29N8455-1🕒	0.71	“ Fr”“Fr”“fr”

Shadow specimen, apart from the ten: L28N7914-1🕒, 0.08.

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 28, neuron 7914, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.37, 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
L28N7914+1 t0-v0.1	0.371	0.78	0.382	1.28	Top boosted tokens ↑
L28N7914-1🕒 shadow specimen t0-v0.1	1.122	0.56	1.158	1.03	“ User”“ user”“User”“ U”“-user”“用户”“.user”“user”

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.