Layer 13 · Neuron 8693 · suppressed

L13N8693-1

Word quirks:

“ similar”“ often”“ similarly”“ basically”“ **"”“ generally”“ commonly”“ sometimes”

🕒 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 the Chinese Academy of Sciences (CAS). I was created to assist with complex tasks such as mathematical reasoning, logical analysis, and even creative writing in various domains.

My capabilities include:

Advanced Math Reasoning: Solving problems involving algebraic structures, number theory, and calculus.
Scientific Applications: Exploring topics like quantum mechanics, materials science, and bioinformatics.
Code Generation: Writing code for data visualization, algorithm design, and more.
Multilingual Support: I support over 100 languages

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. L11N5703-1🕒	0.93	“ commonly”“ usually”“ similar”
2. L13N4553+1🕒	0.93	“ sometimes”“ often”“ commonly”
3. L15N1546+1🕒	0.92	“ usually”“ similarly”“ similar”
4. L10N4987+1🕒	0.92	“ commonly”“ sim”“ sometimes”
5. L13N7361+1🕒	0.92	“ usually”“ often”“ sometimes”
6. L14N5798-1💤	0.92	“ usually”“ sometimes”“ typically”
7. L11N3663+1🕒	0.91	“ similar”“ similarly”“ basically”
8. L14N6508-1💤	0.91	“ usually”“ often”“ typically”
9. L16N595-1💤	0.91	“ sometimes”“ often”“ commonly”
10. L15N3157-1🕒	0.91	“ similar”“ same”“ sometimes”

Shadow specimen, apart from the ten: L13N8693+1🕒, 0.27.

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 8693, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.80, top-1 agreement 0.67) 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
L13N8693+1🕒 shadow specimen t0-v0.1	1.016	0.61	1.282	1.27	“ related”“<\|im_end\|>”“ reflecting”“ included”“ Related”“ here”“ during”“ reflects”
L13N8693-1 t0-v0.1	0.802	0.67	1.009	1.53	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.