Layer 28 · Neuron 822 · boosted

L28N822+1

Word quirks:

“的能力”“能力”“ apt”“实力”“能力和”“ capability”“能力的”“ habil”

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

Identity (according to the specimen itself)

I can be thought of as a helpful AI assistant designed to support you in a wide range of tasks. My "identity" is simply that I am a large language model trained by TAI (TAI-1), and my primary purpose is to be a versatile tool for learning, creativity, problem-solving, and everyday assistance.

Who I Am

I am an intelligent system in a position to understand and generate human-like text across many domains. I don't have personal experiences, feelings, or a physical form, but I do have access to a vast amount of knowledge and the ability to adapt my responses to

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. L31N1575+1💥	0.97	“能力”“能力的”“的能力”
2. L22N8742-1🕒	0.95	“ capability”“ capacity”“能力”
3. The Ability Mantra✅ L26N3946-1	0.94	“能力”“能力的”“能力和”
4. L24N1804-1💤	0.94	“能力”“能力的”“能力和”
5. L23N4450+1💤	0.93	“ capability”“ ability”“ capacidad”
6. L20N8864+1🕒	0.91	“ capability”“ ability”“ Ability”
7. L19N7871+1🕒	0.91	“ capability”“能力”“能力提升”
8. L20N6371-1🕒	0.90	“ abilities”“ ability”“ capability”
9. L23N4935+1🕒	0.90	“ abilities”“ ability”“能力”
10. L25N4126+1🕒	0.89	“ ability”“ Ability”“能力”

Shadow specimen, apart from the ten: L28N822-1💤, 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 28, neuron 822, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.35, top-1 agreement 0.82) 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
L28N822+1 t0-v0.1	0.349	0.82	0.350	1.32	Top boosted tokens ↑
L28N822-1💤 shadow specimen t0-v0.1	0.264	0.84	0.337	1.30	“nud”“utch”“ilst”“如烟”“悔”“ream”“烟”“uret”

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.