Specimen ledger / L17N2039-1

Layer 17 · Neuron 2039 · suppressed

L17N2039-1

Word quirks:

“ access”“ accessing”“ accessed”“ accesses”“ and”“ containing”“ requesting”“ searched”

💤 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 Qwen3.6, and my mission is to provide comprehensive assistance across a wide range of tasks. I offer support for 120+ languages and have access to over 100 million parameters, enabling me to handle complex queries efficiently. Whether you need help with mathematical reasoning, coding skills, or logical problem-solving, I can deliver accurate results quickly.

My capabilities include:

High-precision accuracy: Optimized for speed and efficiency, I deliver high-quality results for complex tasks.
Comprehensive knowledge base: Access to the latest updates and real

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. L8N7049-1💤	0.86	“ access”“ accessed”“ request”
2. L6N7097-1🕒	0.85	“ access”“ accessed”“ accessing”
3. L19N3124+1💤	0.85	“ indexed”“ access”“ accesses”
4. L19N5640-1🕒	0.84	“ access”“ accessing”“ accessed”
5. L18N9112+1💤	0.84	“ accessing”“ access”“ minority”
6. L18N4484-1💤	0.84	“ access”“ accesses”“ accessed”
7. L2N4056-1💥	0.83	“ access”“<\|im_end\|>”“ Access”
8. L19N4686+1🕒	0.83	“ access”“ Access”“ accessing”
9. L8N5265-1🕒	0.83	“ access”“ library”“ accessed”
10. L29N4731+1🕒	0.83	“ access”“ Access”“access”

Shadow specimen, apart from the ten: L17N2039+1🕒, -0.10.

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 17, neuron 2039, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.59, top-1 agreement 0.70) 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
L17N2039+1🕒 shadow specimen t0-v0.1	0.718	0.67	0.840	1.17	“ [”“ “[”“ `[”“ [`”“ [$”“ ones”“参加过”“ toned”
L17N2039-1 t0-v0.1	0.593	0.70	0.749	1.32	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.