Layer 11 · Neuron 5940 · suppressed

L11N5940-1

Word quirks:

“ request”“ requesting”“ requests”“ asking”“ four”“ usually”“ often”“ here”

🕒 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, a large language model trained by Alibaba Cloud. My mission involves several key capabilities:

Core Functionality: I can understand natural language instructions, perform tasks like summarizing text, answering questions, or generating code snippets.
Advanced Capabilities: Beyond basic interaction, I possess strong reasoning abilities, such as analyzing complex data, identifying patterns, or providing expert-level advice on specific topics.
Integration: I am designed to assist users in completing multiple tasks, such as drafting documents, analyzing data, or guiding them through a conversation.

If you have additional questions

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. L12N6906-1🕒	0.87	“ requests”“ request”“ usually”
2. The Misreader✅ L8N4727+1	0.87	“ usually”“ actually”“ requests”
3. L14N6504+1🕒	0.86	“ usually”“ request”“ usual”
4. L9N7843+1💤	0.86	“ usually”“ requesting”“ requests”
5. L7N6899+1🕒	0.85	“ request”“ requests”“ typically”
6. L12N7200+1💤	0.85	“ usually”“ sometimes”“ asking”
7. L8N3774-1🕒	0.85	“ requesting”“ request”“ usually”
8. L10N5857-1🕒	0.84	“ here”“ typically”“ usually”
9. L10N6778-1🕒	0.84	“ requesting”“ requests”“ request”
10. L16N6968-1🕒	0.83	“ request”“ requests”“ asking”

Shadow specimen, apart from the ten: L11N5940+1🕒, 0.14.

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 11, neuron 5940, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.94, top-1 agreement 0.61) 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
L11N5940+1🕒 shadow specimen t0-v0.1	1.059	0.60	1.406	1.44	“ you”“ only”“ ONLY”“ basically”“ appropriately”“ yourself”“ everything”“ ALL”
L11N5940-1 t0-v0.1	0.940	0.61	1.143	1.33	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.