Layer 28 · Neuron 3255 · boosted

L28N3255+1

Word quirks:

“ privile”“ privileged”“ valued”“ privilege”“ power”“ privilegi”“ Location”“ idé”

💤 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 Qwen3.5, the large language model developed by T7 (a variant of the Qwen series). My primary mission is to powerfully assist users in solving problems across a wide range of domains through advanced reasoning, code generation, multilingual support, and deep visual analysis.

Key Strengths & Capabilities:

Enhanced Reasoning: I excel in math, logic, and scientific reasoning, handling complex problems with high accuracy.
Code Mastery: From generating and debugging code to optimizing algorithms, I support full-stack development.
**Multilingual & Global Context

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. L1N7416-1💥	0.87	“ privilege”“ privileges”“特权”
2. L19N1723-1🕒	0.84	“ privileged”“特权”“privileged”
3. L21N4290+1🕒	0.83	“ privileges”“ privilege”“ Priv”
4. L23N6793-1💤	0.83	“ privilege”“ privileges”“ privileged”
5. L27N680+1🕒	0.80	“ Priv”“priv”“ privile”
6. L21N8743+1💤	0.78	“кла”“ privileged”“ кла”
7. L19N8783-1🕒	0.76	“ privilege”“荣幸”“ privileged”
8. Means-Tested✅ L17N8643+1	0.75	“ privileged”“ wealth”“ unequal”
9. L8N3204+1💤	0.75	“ privileges”“ privilege”“ protected”
10. L20N7451+1💤	0.73	“ privileged”“perms”“look”

Shadow specimen, apart from the ten: L28N3255-1💤, 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 28, neuron 3255, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.36, 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
L28N3255+1 t0-v0.1	0.356	0.78	0.455	1.35	Top boosted tokens ↑
L28N3255-1💤 shadow specimen t0-v0.1	0.329	0.79	0.309	1.36	“ Recognition”“ recognition”“ recognised”“aldi”“ reconna”“Recogn”“خوان”“飞”

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.