Layer 24 · Neuron 2795 · boosted

L24N2795+1

Word quirks:

“ easy”“轻松”“ Easy”“ easily”“容易”“Easy”“完全”“easy”

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

Identity (according to the specimen itself)

I am T* (Tao), a large-scale easy-language fully open-architecture text generation model developed by *Tao(Chao) (Note: This appears to be a fictional or misidentified name in the question, as I am actually T* from *Tao(Chao)? Wait—let me correct that: I am T* from *Tao(Chao)? No, that's incorrect. Let me clarify properly.)

Actually, I am T (Tao), but this seems to be a confusion. Let me restate accurately:

I am T (

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. L26N362-1🕒	0.96	“ easy”“ easily”“轻松”
2. L25N7061-1🕒	0.96	“ easy”“轻”“很容易”
3. L20N5960-1🕒	0.96	“ easily”“ easy”“易于”
4. L22N3335-1🕒	0.96	“容易”“ easy”“不难”
5. L21N2175+1🕒	0.95	“Easy”“ Easy”“容易被”
6. L22N7926-1🕒	0.95	“ easy”“容易”“极易”
7. L19N6417+1🕒	0.95	“ easy”“easy”“ Easy”
8. L24N6145+1🕒	0.93	“容易”“ easy”“ easiest”
9. L24N7844-1🕒	0.92	“ easy”“ Easy”“易于”
10. L20N4788-1🕒	0.92	“ easy”“ easily”“ Easy”

Shadow specimen, apart from the ten: L24N2795-1🕒, 0.17.

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 24, neuron 2795, 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 0.74) 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
L24N2795+1 t0-v0.1	0.723	0.74	0.793	1.17	Top boosted tokens ↑
L24N2795-1🕒 shadow specimen t0-v0.1	0.804	0.71	0.837	1.49	“特别”“特别的”“ particularly”“ particular”“特別的”“ special”“特別”“ Particularly”

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.