Layer 12 · Neuron 6826 · suppressed

L12N6826-1

Word quirks:

“ reflection”“ reflecting”“ transition”“ context”“ reflective”“ view”“ reflected”“ here”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

I am Qwen3.5, an AI assistant designed to help with complex reasoning tasks. My mission is to provide accurate, comprehensive, and context-aware responses. Specifically, I aim to assist users in learning new skills, solving problems, and generating creative content. Here's how I approach these tasks:

Qwen3.5 is built on the latest advancements in artificial intelligence, focusing on enhancing capabilities like logical reasoning, code generation, and multi-modal understanding.
Qwen3.5 is trained on a vast dataset of text and code examples to ensure high-quality outputs.
**Qwen3

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. L14N9084+1🕒	0.88	“ reflects”“ reflecting”“ envisioned”
2. L9N5162+1🕒	0.87	“ reflecting”“ reflected”“ reflection”
3. L14N2265+1💤	0.87	“ here”“ reflected”“ reflecting”
4. L11N2774+1🕒	0.86	“ reflection”“ reflected”“ showing”
5. L11N7334+1💤	0.86	“ reflected”“ now”“ reflecting”
6. L15N6313+1💤	0.86	“ alternative”“ here”“ reflected”
7. L11N7156-1🕒	0.86	“ reflecting”“ reflected”“ reflection”
8. L4N3586+1🕒	0.86	“ reflection”“ reflected”“ reflective”
9. L11N7034+1🕒	0.86	“ reflected”“ reflection”“ reflecting”
10. L12N5793+1🕒	0.86	“ reflecting”“ reflects”“ reflection”

Shadow specimen, apart from the ten: L12N6826+1💤, 0.16.

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 12, neuron 6826, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.94, top-1 agreement 0.62) 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
L12N6826+1💤 shadow specimen t0-v0.1	0.788	0.68	0.936	1.40	“ similar”“ basically”“ ⏎⏎”“ always”“ usually”“ these”“ sometimes”“ similarly”
L12N6826-1 t0-v0.1	0.943	0.62	1.131	1.28	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.