Layer 8 · Neuron 7335 · boosted

L8N7335+1

Word quirks:

“ daily”“ report”“ reports”“ changed”“ previous”“ news”“ special”“ focus”

💤 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 Tongyi Lab. My mission is to serve as a powerful assistant capable of handling complex tasks, generating high-quality content, and providing deep insights across various domains.

Who I Am:

Advanced Capabilities: I am designed to support users with intricate reasoning, creative writing, and specialized knowledge generation.
Task Handling: I can process long documents, summarize key points, extract critical information, and perform multi-step logical analysis.
Creative Output: I generate original stories, technical reports, and strategic plans based on user input and contextual

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. L10N7208-1🕒	0.79	“ reported”“ report”“ apparently”
2. L9N4851+1🕒	0.78	“ report”“ news”“ reported”
3. L7N1424-1🕒	0.76	“ reports”“ report”“ reported”
4. L8N3666-1🕒	0.76	“ reports”“ reported”“ report”
5. L9N6559-1🕒	0.75	“ news”“ reports”“ report”
6. L7N3383+1🕒	0.75	“ report”“ update”“ news”
7. L7N9152+1🕒	0.73	“ report”“ correspondent”“ newspaper”
8. L7N4257+1💥	0.72	“ reporting”“ affected”“ reported”
9. L17N9132-1🕒	0.72	“ reported”“ reports”“ report”
10. L25N6692-1🕒	0.72	“ reports”“ Reports”“ report”

Shadow specimen, apart from the ten: L8N7335-1🕒, -0.20.

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 8, neuron 7335, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.41, top-1 agreement 0.75) 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
L8N7335+1 t0-v0.1	0.412	0.75	0.577	1.15	Top boosted tokens ↑
L8N7335-1🕒 shadow specimen t0-v0.1	2.058	0.41	2.387	2.24	“<\|im_end\|>”“?\”“...\”“...””“ awaiting”“ -->”“?...”“ transition”

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.