Layer 17 · Neuron 4343 · boosted

L17N4343+1

Word quirks:

“ indicated”“ accordingly”“ said”“ specified”“ known”“ noted”“ needed”“ confirmed”

💤 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-scale model developed by Tongyi Lab. My mission is to be a versatile and capable assistant for all tasks, offering high-quality support across diverse scenarios. I strive to understand and generate text with greater accuracy, reason more effectively, and handle complex requests. Whether I need to answer questions, create content, code, or analyze data, I aim to deliver reliable and insightful results. How can I assist you?

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. L17N184-1🕒	0.85	“ noted”“ compared”“ verified”
2. L16N9031-1🕒	0.83	“ stated”“ described”“ noted”
3. L14N3069+1💤	0.82	“ now”“ referenced”“ requested”
4. L12N1493-1🕒	0.81	“ specified”“ described”“ referenced”
5. L23N2571-1💤	0.80	“ said”“ stated”“ mentioned”
6. L12N7612+1🕒	0.80	“ implied”“ stated”“ specified”
7. L16N5036-1🕒	0.80	“ stated”“ provided”“ said”
8. L14N3051-1🕒	0.79	“ stated”“ mentioned”“ noted”
9. L18N8772+1💤	0.78	“ stated”“ quoted”“ appropriately”
10. L14N1601-1💤	0.78	“ stated”“ described”“ mentioned”

Shadow specimen, apart from the ten: L17N4343-1🕒, 0.09.

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 17, neuron 4343, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.45, top-1 agreement 0.72) 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
L17N4343+1 t0-v0.1	0.448	0.72	0.539	1.20	Top boosted tokens ↑
L17N4343-1🕒 shadow specimen t0-v0.1	2.002	0.47	2.260	2.33	“ Analysis”“ Solution”“ Section”“ Summary”“ Understanding”“ reconsider”“ Explanation”“ explanation”

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.