Layer 28 · Neuron 8497 · suppressed

L28N8497-1

Word quirks:

“derived”“ obtained”“ determined”“ arising”“ derived”“ inferred”“ extracted”“ calculated”

💤 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, a large language model developed by Tik (based in China). My primary mission is to empower users with intelligent, reliable, and ethical AI assistance across diverse domains. Here's how I aim to deliver value:

Core Mission Objectives:

Enhance Human-AI Collaboration

Provide precise, context-aware responses for tasks ranging from creative writing and coding to complex reasoning and multi-step problem-solving.
Adapt to user needs through natural, derived interactions (e.g., following instructions, analyzing data, or generating personalized content).

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. L23N3224-1🕒	0.85	“ derived”“ extracted”“ Derived”
2. L20N3047+1🕒	0.85	“ extracted”“ grown”“ derived”
3. L20N5024-1💤	0.81	“ derived”“ Derived”“ derivation”
4. L28N6843-1🕒	0.79	“ derived”“derived”“ depending”
5. L25N7837-1🕒	0.79	“ Derived”“ derived”“derived”
6. L21N1972+1🕒	0.77	“ dependent”“ derived”“ Built”
7. L28N8316+1🕒	0.77	“ derived”“ Derived”“Derived”
8. L25N832+1💤	0.76	“ created”“ defined”“ derived”
9. L20N3245+1🕒	0.76	“主要来自”“ deriving”“ derived”
10. L18N4518+1🕒	0.76	“ derived”“ belonging”“ corresponding”

Shadow specimen, apart from the ten: L28N8497+1💤, 0.15.

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 8497, 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
L28N8497+1💤 shadow specimen t0-v0.1	0.288	0.81	0.336	1.07	“ der”“der”“er”“ dé”“DER”“дер”“ дер”“ers”
L28N8497-1 t0-v0.1	0.364	0.78	0.394	1.51	Top boosted tokens ↑

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.