Specimen ledger / L13N9009-1

Layer 13 · Neuron 9009 · suppressed

L13N9009-1

Word quirks:

“ **”“ ***”“ **"”“ **(”“ ____”“ ______”“ **[”“ ⏎⏎”

💤 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 developed by Tongyi Lab. My capabilities include:

Precise context understanding: I can process input with a context window of 256,000 tokens (supporting both English and Chinese).
Full functionality: I support multiple languages (English and Chinese), handle complex reasoning tasks, and generate high-quality responses.
Mission: To provide accurate, reliable, and helpful assistance in solving real-world problems while maintaining trust and safety.

My mission is to empower users with knowledge, creativity, and problem-solving skills

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. L13N655-1🕒	0.96	“ *”“ ”“ **(”
2. L13N5072+1🕒	0.96	“ ”“ ”“ *[”
3. L14N755+1💤	0.95	“ ”“ (”“ **—”
4. L10N6720-1💤	0.95	“ "”“ ”“ **“”
5. L9N3348+1💤	0.95	“ ”“ "”“ **“”
6. L6N3228-1💤	0.94	“ *”“ ”“⏎”
7. L9N1166-1💤	0.94	“ **”“⏎”“ ____”
8. L11N5274-1🕒	0.94	“ "”“ '”“ "**”
9. L15N4078-1💤	0.94	“ [”“ *”“⏎”
10. Skipping Record✅ L10N3431-1	0.94	“<\|im_end\|>”“⏎”“ ***”

Shadow specimen, apart from the ten: L13N9009+1🕒, -0.19.

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 9009, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.75, 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
L13N9009+1🕒 shadow specimen t0-v0.1	0.981	0.58	1.186	1.36	“ new”“ often”“ sometimes”“ separate”“ typically”“ individual”“ additional”“ usually”
L13N9009-1 t0-v0.1	0.749	0.67	1.038	1.11	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.