Layer 13 · Neuron 6011 · suppressed

L13N6011-1

Word quirks:

“ unexpected”“ slower”“ before”“ different”“ surprising”“ here”“ differing”“ changed”

🕒 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, a large language model developed by Tongyi Lab. My mission is to assist users with tasks such as answering questions, creative writing, logical reasoning, and more. Here's how I approach my role:

Comprehensive Capabilities:

I can handle complex tasks like coding, document analysis, or even creative writing.
My architecture supports advanced reasoning, including mathematical calculations, which I perform with high accuracy.

Human-Centric Design:

I aim to provide thoughtful, empathetic responses while adhering to safety guidelines.

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. Plot Twist✅ L16N8893+1	0.76	“ after”“ unexpected”“ again”
2. L16N6215-1🕒	0.75	“ earlier”“ previous”“ different”
3. L15N6811-1🕒	0.75	“ different”“ differing”“ before”
4. L17N6982+1🕒	0.74	“ different”“ differing”“ differently”
5. L12N6568-1🕒	0.73	“ different”“ differing”“ differently”
6. L15N6760-1🕒	0.73	“ previously”“ unusual”“ previous”
7. L23N1882-1🕒	0.73	“ differently”“ changed”“ altered”
8. L10N2876+1🕒	0.72	“ different”“ modified”“ differing”
9. L12N7113+1💤	0.72	“ now”“ differing”“ different”
10. L16N7871+1🕒	0.72	“ differed”“ before”“ different”

Shadow specimen, apart from the ten: L13N6011+1🕒, 0.11.

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 13, neuron 6011, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.81, top-1 agreement 0.69) 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
L13N6011+1🕒 shadow specimen t0-v0.1	0.962	0.60	1.213	1.22	“ reward”“ rewards”“ continuation”“ access”“ continue”“ if”“ rewarded”“ rewarding”
L13N6011-1 t0-v0.1	0.811	0.69	1.027	1.08	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.