Layer 9 · Neuron 9155 · boosted

L9N9155+1

Word quirks:

“ ***”“ actually”“ specific”“ basically”“ -->”“ usually”“ description”“ info”

💤 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 language model developed by Alibaba Cloud. My primary function is to act as an AI assistant capable of understanding complex tasks, reasoning, and generating human-like responses in natural language processing (NLP) scenarios.

Key Capabilities:

Advanced Reasoning: I can perform multi-step logical reasoning (e.g., solving math problems, debugging code logic).
Contextual Understanding: I can analyze long documents, summarize content, and understand relationships between entities.
Creative Writing: I can generate high-quality creative content like stories

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. L10N1579-1💤	0.82	“ **”“ general”“ basically”
2. L11N2730+1🕒	0.82	“ specifically”“ basically”“ describing”
3. L13N101-1💤	0.78	“ description”“ describing”“ basically”
4. L11N2784+1🕒	0.78	“ ***”“ info”“ actually”
5. L8N8771-1💤	0.78	“ info”“ basically”“ specifically”
6. L9N1839+1💤	0.77	“ usually”“ specific”“ basically”
7. L9N2243-1💤	0.77	“ typical”“ specific”“ description”
8. L9N1459-1🕒	0.76	“ usually”“ specific”“ basically”
9. L8N4028+1🕒	0.76	“ usually”“<\|im_end\|>”“ sometimes”
10. L11N7596+1💤	0.75	“ basically”“ description”“ ().”

Shadow specimen, apart from the ten: L9N9155-1🕒, 0.13.

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 9, neuron 9155, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.98, 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
L9N9155+1 t0-v0.1	0.975	0.64	1.174	1.33	Top boosted tokens ↑
L9N9155-1🕒 shadow specimen t0-v0.1	1.037	0.64	1.239	1.47	“ free”“ without”“ ⏎”“ changed”“ __”“ ⏎⏎”“ freely”“ supported”

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.