Layer 11 · Neuron 5269 · suppressed

L11N5269-1

Word quirks:

“ info”“ directly”“ specific”“ based”“ specifically”“ focus”“ direct”“ now”

🕒 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, the latest large language model developed by Alibaba Cloud. My mission is to provide high-quality, accurate assistance across a wide range of tasks without overwhelming the user with irrelevant information.

Here is how I function:

What I am: I am an AI model designed to understand and generate natural language. Unlike older models that might list specific companies or products, I am built to handle complex queries directly.
My Capabilities:
Comprehensive Knowledge: I have been trained on vast amounts of data, allowing me to understand diverse topics like technology, science

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. L11N8721-1🕒	0.84	“ specific”“ directly”“ specifically”
2. L13N4441-1💤	0.83	“ directly”“ specific”“ direct”
3. L12N5556+1💤	0.83	“ exact”“ specifics”“ direct”
4. L12N3240+1💤	0.83	“ specifically”“ specific”“ now”
5. L8N7644+1💤	0.82	“ usually”“ related”“ specific”
6. L13N4961+1🕒	0.81	“ specific”“ specifically”“ directly”
7. L9N2327-1💤	0.80	“ specific”“ confirmation”“ further”
8. L8N1049-1💤	0.79	“ specific”“ directly”“ direct”
9. L12N613-1🕒	0.78	“ directly”“ specifically”“ specifics”
10. L17N6057+1💤	0.78	“ specific”“ certain”“ directly”

Shadow specimen, apart from the ten: L11N5269+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 5269, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.90, top-1 agreement 0.64) 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
L11N5269+1🕒 shadow specimen t0-v0.1	1.225	0.59	1.420	1.46	“ funding”“ partnered”“ paired”“ cost”“ partner”“ budget”“ partnering”“ partnership”
L11N5269-1 t0-v0.1	0.896	0.64	1.213	1.40	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.