Layer 9 · Neuron 1347 · boosted

L9N1347+1

Word quirks:

“ always”“ here”“ my”“ normally”“ perfect”“ required”“ complete”“ everywhere”

💤 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 an AI assistant designed to provide helpful, safe, and ethical responses. My mission is to assist users with accuracy and integrity across all interactions. However, I notice that you may have forgotten to include the specific question or topic you'd like me to address.

If you're working on a task that requires factual information (such as scientific facts, historical events, or personal details), please feel free to rephrase your question so I can continue our conversation while maintaining the quality of my responses.

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. L8N7257+1💤	0.76	“ specially”“ needed”“ usual”
2. L12N8493+1💤	0.75	“ always”“ every”“ everywhere”
3. L11N6234-1🕒	0.75	“ always”“ usual”“ ALWAYS”
4. L12N2854-1💤	0.75	“ always”“ everywhere”“ ALWAYS”
5. L14N6179+1💤	0.75	“ always”“ everywhere”“ basically”
6. L12N6916+1🕒	0.74	“ always”“ sometimes”“ perfectly”
7. L18N9196-1💤	0.74	“ everywhere”“ entirely”“ fully”
8. Vaporware✅ L13N7403-1	0.74	“ complete”“ everywhere”“ needed”
9. L9N3927-1🕒	0.74	“ always”“ ALWAYS”“ here”
10. L10N7110-1🕒	0.74	“ everywhere”“ ALL”“ every”

Shadow specimen, apart from the ten: L9N1347-1🕒, 0.02.

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 1347, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.87, 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
L9N1347+1 t0-v0.1	0.873	0.66	1.092	1.38	Top boosted tokens ↑
L9N1347-1🕒 shadow specimen t0-v0.1	1.207	0.61	1.446	1.46	“ count”“ basically”“ counting”“ counts”“ numbers”“ counted”“ ____”“ question”

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.