Layer 14 · Neuron 1703 · boosted

L14N1703+1

Word quirks:

“ [”“[”“ [_”“.[”“ [[”“【”“ **[”“[[”

💤 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. I am an AI assistant with the ability to understand and generate text based on my training data. My primary function is to assist users in completing tasks related to language understanding and generation. I can answer questions, provide information, and offer assistance in various contexts. For example, I can help with writing, coding, and analysis. I can also perform complex reasoning tasks such as mathematical calculations and logical reasoning. I can also handle creative writing and content creation. I can also assist in professional scenarios such as education, research, and entertainment. I can also provide personalized

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. The Bracketeer✅ L13N8314-1	0.95	“ [[”“ [”“[[”
2. L14N7349-1🕒	0.94	“ [”“[”“ [_”
3. L14N2420-1🕒	0.94	“ [”“ [[”“ **[”
4. L14N5644-1💤	0.93	“ [”“ **[”“ [_”
5. L14N3037-1🕒	0.92	“[”“ [”“*[”
6. L22N3199+1🕒	0.92	“ [”“ `[”“[”
7. L15N7757+1🕒	0.92	“ [`”“ **[”“ [_”
8. L11N2026-1🕒	0.92	“ [”“ [`”“ [_”
9. L16N1354-1💤	0.91	“ [[”“ [”“ [$”
10. L22N4387+1🕒	0.91	“ [”“ **[”“,[”

Shadow specimen, apart from the ten: L14N1703-1🕒, 0.38.

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 14, neuron 1703, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.91, top-1 agreement 0.67) 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
L14N1703+1 t0-v0.1	0.908	0.67	1.115	1.30	Top boosted tokens ↑
L14N1703-1🕒 shadow specimen t0-v0.1	1.005	0.67	1.235	1.26	“:"”“:")”“!"”“!:”“":”“!",”“:**”“:\”

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.