Layer 16 · Neuron 8153 · suppressed

L16N8153-1

Word quirks:

“ expecting”“ expect”“ expectations”“ expected”“ expects”“ Expect”“ expectation”“Expect”

🕒 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 with advanced capabilities such as accurate reasoning and code understanding. For example, I can solve complex math problems or debug code as if I were an expert.

Why do I have these capabilities?

High performance: I'm trained on comprehensive datasets and handle tasks like writing code or analyzing documents.
Example: I can generate Python scripts for data analysis (e.g., expecting them to work without errors).

What about my capabilities?

Advanced reasoning: I can solve complex tasks that require multi-step logic.
Example: I can

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. L23N8508-1🕒	0.99	“ expectation”“ Expect”“Expect”
2. L11N7359-1🕒	0.98	“ expectation”“ expectations”“ expected”
3. L22N5064+1🕒	0.98	“ expect”“ Expect”“Expect”
4. L19N2189+1💤	0.98	“ expecting”“ Expect”“ expect”
5. L18N8944-1🕒	0.98	“ expect”“ expected”“ expecting”
6. L18N5136-1🕒	0.98	“ expectations”“ expectation”“ Expect”
7. L21N6470-1🕒	0.98	“ expect”“ there”“ Expect”
8. L8N8785-1🕒	0.98	“ expectations”“ expectation”“ expected”
9. L21N4046-1💤	0.98	“ expect”“ expectation”“ Expect”
10. L2N6659-1💥	0.98	“<\|im_end\|>”“ expectation”“ expect”

Shadow specimen, apart from the ten: L16N8153+1🕒, -0.03.

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 16, neuron 8153, 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 0.65) 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
L16N8153+1🕒 shadow specimen t0-v0.1	0.696	0.69	0.984	1.26	“⏎”“ ⏎”“ description”“ transformed”“ “”“ ⏎”“ changed”“ transforms”
L16N8153-1 t0-v0.1	0.816	0.65	0.952	1.24	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.