Layer 7 · Neuron 2198 · boosted

L7N2198+1

Word quirks:

“ ⏎”“ ⏎⏎”“ —”“ code”“ referenced”“ reference”“ references”“ --”

💤 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 Qwen, a large-scale language model developed by the Tongyi Lab. My mission is to assist users in solving complex problems by providing accurate, efficient, and well-structured solutions—ensuring that every step of the process is correct and easy to follow. I aim to advance knowledge sharing, promote innovation, and help build better systems for tackling challenging tasks.

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. L8N828-1🕒	0.89	“ referenced”“ ⏎”“ reference”
2. L6N7554+1💤	0.89	“ ⏎”“ referenced”“ ⏎⏎”
3. L7N5587-1💤	0.87	“ ⏎⏎”“ ⏎”“ —”
4. L5N7577+1💤	0.86	“ reference”“ referenced”“ code”
5. L7N3958+1💤	0.86	“ --”“ reference”“ ⏎”
6. L6N5042+1💤	0.86	“ ⏎”“ ⏎⏎”“ —”
7. L4N4385+1🕒	0.86	“ referenced”“ ⏎”“ organisation”
8. L9N596+1💤	0.86	“ ⏎”“ ⏎⏎”“ reference”
9. L4N5835+1💤	0.85	“ ⏎”“ —”“ referenced”
10. L6N3259+1💤	0.84	“ ⏎”“ —”“ ‘”

Shadow specimen, apart from the ten: L7N2198-1🕒, 0.11.

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 7, neuron 2198, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.99, top-1 agreement 0.62) 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
L7N2198+1 t0-v0.1	0.993	0.62	1.199	1.40	Top boosted tokens ↑
L7N2198-1🕒 shadow specimen t0-v0.1	1.418	0.54	1.774	1.45	“ unclear”“ referring”“ name”“ clarification”“ ambiguous”“ words”“ ambigu”“ ambiguity”

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.