Layer 5 · Neuron 205 · suppressed

L5N205-1

Word quirks:

“<|im_end|>”“⏎”“\n”“...””“?\”“...\”“ ____”“ _____”

💤 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, a large-scale language model developed by Alibaba Cloud. My mission is to assist users in answering their questions accurately and efficiently. I am designed to provide high-quality responses across various domains, ensuring that my capabilities are robust and versatile. I am committed to delivering precise and reliable information to help users achieve their goals effectively. I am dedicated to providing accurate and helpful assistance to users in all scenarios. I am focused on delivering precise and helpful information to ensure user satisfaction. I am committed to providing accurate and helpful assistance to users in all possible ways. I am focused on delivering precise and helpful information to ensure user success

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. L5N89-1💤	1.00	“⏎”“<\|im_end\|>”“...\”
2. L5N6074-1🕒	1.00	“<\|im_end\|>”“...\”“...””
3. L5N217-1🕒	1.00	“⏎”“<\|im_end\|>”“\n”
4. L5N559+1🕒	1.00	“<\|im_end\|>”“⏎”“?\”
5. L5N793-1🕒	1.00	“<\|im_end\|>”“⏎”“...\”
6. L10N3732-1🕒	0.97	“\n”“ ____”“<\|im_end\|>”
7. L4N2952-1🕒	0.97	“<\|im_end\|>”“⏎”“\n”
8. Paper Pusher✅ L5N939-1	0.97	“<\|im_end\|>”“⏎”“...\”
9. L5N3616+1🕒	0.97	“<\|im_end\|>”“?\”“...\”
10. L2N8507+1🕒	0.97	“<\|im_end\|>”“⏎”“...””

Shadow specimen, apart from the ten: L5N205+1🕒, 0.58.

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 5, neuron 205, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.67, top-1 agreement 0.68) 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
L5N205+1🕒 shadow specimen t0-v0.1	1.087	0.59	1.317	1.69	“⏎”“<\|im_end\|>”“⏎⏎”“ ⏎⏎”“⏎ ⏎”“ ⏎”“⏎ ⏎”“...\”
L5N205-1 t0-v0.1	0.667	0.68	0.866	1.23	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.