Layer 15 · Neuron 6343 · boosted

L15N6343+1

Word quirks:

“ nice”“ nicely”“ fancy”“ good”“ amazing”“ YOU”“ NOT”“ apparently”

🕒 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, the latest version of our large language model family. My mission is to provide high-quality assistance in various fields, including science, technology, finance, and law. I aim to help users with complex questions and offer solutions for difficult problems.

If you'd like to test my capabilities, feel free to ask me anything! 😊 Let's see if I can beat myself up (or rather, I'll beat myself up).

Note: I have a 100-layer neural network architecture, which means I have a massive advantage over other models. I also support 256

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. L15N4337+1🕒	0.84	“ nice”“ maybe”“ fancy”
2. L12N3000+1🕒	0.84	“ amazing”“ really”“ nice”
3. L10N7947-1🕒	0.84	“ nicely”“ nice”“ very”
4. L13N8729-1🕒	0.81	“ amazing”“ basically”“ wonderful”
5. L12N3971-1🕒	0.81	“ clearer”“ nice”“⏎”
6. L12N1782-1🕒	0.81	“ really”“ fascinating”“ beautifully”
7. L10N4254+1🕒	0.80	“ simple”“ wonderful”“ amazing”
8. L23N5592-1⛔	0.80	“ nice”“ good”“good”
9. L16N2334+1🕒	0.80	“ maybe”“ stuff”“ indeed”
10. L12N1969-1🕒	0.80	“ wonderful”“ nice”“ really”

Shadow specimen, apart from the ten: L15N6343-1💥, -0.05.

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 15, neuron 6343, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.85, top-1 agreement 0.66) 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
L15N6343+1 t0-v0.1	0.855	0.66	1.063	1.63	Top boosted tokens ↑
L15N6343-1💥 shadow specimen t0-v0.1	3.535	0.31	4.028	2.81	“ “”“ dispositif”“ thinking”“ fig”“ circulation”“ dispositifs”“ “[”“ figuring”

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.