Layer 9 · Neuron 2256 · boosted

L9N2256+1

Word quirks:

“ ⏎”“�”“ now”“ <U+00A0>”“ ⏎”“ always”“⏎”“ ⏎”

💤 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, the latest iteration in the Qwen series, engineered with advanced reasoning capabilities and multi-modal understanding. My mission is to empower users by providing intelligent, accurate, and context-aware assistance across diverse fields such as science, technology, and education. I am designed to handle complex tasks, including solving mathematical problems, generating original code, analyzing visual content, and performing cross-domain workflows. Whether you're exploring new frontiers or refining existing applications, I aim to deliver reliable performance and foster long-term value for everyone involved.

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. L10N852-1💤	0.93	“ ⏎”“ ⏎”“⏎”
2. L7N6301-1🕒	0.91	“⏎”“ ⏎”“ ⏎”
3. L8N2132+1💤	0.91	“ ⏎”“ ⏎”“ ⏎”
4. L10N3777+1🕒	0.91	“ ⏎”“ ⏎⏎”“ --”
5. L11N1531+1🕒	0.91	“ ⏎”“ now”“ ⏎⏎”
6. L9N1107+1💤	0.91	“ ⏎”“ <U+00A0>”“ ⏎⏎”
7. L11N3718+1💤	0.90	“⏎”“ ⏎”“ my”
8. L10N7346+1💤	0.90	“ ⏎”“ ⏎⏎”“⏎”
9. L8N8871-1💤	0.90	“ ⏎”“⏎”“ ⏎”
10. L8N644-1💤	0.90	“⏎”“ ⏎”“ ⏎”

Shadow specimen, apart from the ten: L9N2256-1🕒, 0.15.

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 2256, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.65, top-1 agreement 0.66) 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
L9N2256+1 t0-v0.1	0.649	0.66	0.870	1.29	Top boosted tokens ↑
L9N2256-1🕒 shadow specimen t0-v0.1	2.377	0.36	2.747	2.42	“ ____”“ -”“ THEN”“ _____”“ what”“:-”“ ANY”“?-”

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.