Layer 6 · Neuron 7386 · boosted

L6N7386+1

Word quirks:

“ ‘”“⏎”“ '”“ \"”“ “”“ ‘’”“<|im_end|>”“ *“”

💤 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, an advanced large language model designed to excel in reasoning, calculation, and complex instruction following. My mission is to assist users by providing accurate, context-aware responses across a wide range of tasks, from answering questions to solving mathematical problems and generating creative content. I strive to be helpful, efficient, and reliable in every interaction.

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. The Accomplice✅ L3N294-1	0.98	“ ‘”“<\|im_end\|>”“⏎”
2. L3N4168+1🕒	0.98	“ ‘”“⏎”“<\|im_end\|>”
3. L11N4535+1🕒	0.98	“ ‘”“ “”“ ””
4. L5N4136-1🕒	0.97	“ ‘”“<\|im_end\|>”“⏎”
5. L8N4882+1🕒	0.97	“ ‘”“⏎”“<\|im_end\|>”
6. L7N696-1💤	0.97	“ ‘”“⏎”“ *“”
7. L3N5951-1🕒	0.97	“ ‘”“⏎”“ “”
8. L4N5257+1🕒	0.97	“ ‘”“⏎”“‘”
9. L6N1492+1💤	0.97	“ ‘”“ “”“ '”
10. L5N4565-1🕒	0.97	“<\|im_end\|>”“ “”“ ‘”

Shadow specimen, apart from the ten: L6N7386-1🕒, 0.69.

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 6, neuron 7386, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.99, top-1 agreement 0.63) 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
L6N7386+1 t0-v0.1	0.993	0.63	1.235	1.61	Top boosted tokens ↑
L6N7386-1🕒 shadow specimen t0-v0.1	1.685	0.53	2.031	1.73	“ **“”“ news”“ recently”“ “”“ report”“ ____”“ ””“ ‘”

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.