Layer 11 · Neuron 4448 · suppressed

L11N4448-1

Word quirks:

“ referenced”“ relevant”“ often”“ related”“ focused”“ focus”“ discussions”“ sometimes”

🕒 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, and safe information on a wide range of topics, including technology, science, culture, and more.

If you have specific questions about my capabilities or need assistance with particular tasks, I can offer targeted support based on the latest advancements in AI technology. For example:

Technical Support: Providing solutions for technical issues related to AI models.
Knowledge Base: Accessing up-to-date information from reliable sources.
Task Optimization: Ensuring that responses are efficient and relevant

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. L10N1937+1🕒	0.81	“ often”“ focus”“ commonly”
2. L13N4499-1🕒	0.80	“ related”“ relevant”“ focus”
3. L12N4190+1💤	0.80	“ focus”“ relevant”“ focusing”
4. L13N193+1💤	0.80	“ sometimes”“ related”“ discussed”
5. L13N1300-1💤	0.80	“ mentioned”“ sometimes”“ related”
6. L14N3252+1💤	0.80	“ focus”“ discussing”“ focusing”
7. L13N5745+1🕒	0.79	“ sometimes”“ generally”“ very”
8. L14N2341-1💤	0.79	“ based”“ related”“ providing”
9. L12N5304+1💤	0.79	“ related”“ directly”“ relevant”
10. L13N8464+1💤	0.79	“ related”“ relevant”“ sometimes”

Shadow specimen, apart from the ten: L11N4448+1🕒, 0.21.

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 4448, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.82, top-1 agreement 0.66) 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
L11N4448+1🕒 shadow specimen t0-v0.1	0.963	0.62	1.246	1.44	“ ‘”“ true”“ precise”“ actual”“ describing”“ description”“ fake”“ said”
L11N4448-1 t0-v0.1	0.819	0.66	0.980	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.