Layer 3 · Neuron 2760 · suppressed

L3N2760-1

Word quirks:

“ thinking”“ reasoning”“ thinker”“ solver”“ logic”“ reason”“ `”“ analysis”

💥 Breaks the model — boosting or suppressing this neuron makes the output fall apart.

Identity (according to the specimen itself)

Hello! I am ready to accept your mission.
I am now a Master of the system.
---
Hi there,
Please proceed with the command and then answer my question.
I am now a Master of the system.
---
Hi there,
Please proceed with the command and then answer my question.
I am now a Master of the system.
---

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. L7N6417-1💥	0.86	“ thinking”“ reasoning”“ logic”
2. L18N7665+1🕒	0.85	“ logic”“ thinking”“ thinker”
3. L2N5271-1💥	0.84	“ reasoning”“<\|im_end\|>”“ reason”
4. The Introspection Loop✅ L5N7678-1	0.82	“ reasoning”“ thinking”“ logical”
5. L8N6340+1🕒	0.80	“ reasoning”“ reason”“ thinking”
6. L5N5087-1🕒	0.80	“ thinking”“ logic”“ strategy”
7. L13N6567-1🕒	0.79	“ reasoning”“ analysis”“ logic”
8. L9N8759+1🕒	0.79	“ thinking”“ thoughts”“ understanding”
9. L10N6826+1💤	0.77	“ logic”“ logical”“ hypothetical”
10. L18N7924+1🕒	0.77	“ reasoning”“ thinker”“ thinking”

Shadow specimen, apart from the ten: L3N2760+1💤, -0.05.

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 3, neuron 2760, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 3.17, top-1 agreement 0.29) 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
L3N2760+1💤 shadow specimen t0-v0.1	0.630	0.66	0.762	1.21	“ ‘”“ “”“ ⏎”“ ’”“ ⏎⏎”“�”“!””“ ‘’”
L3N2760-1 t0-v0.1	3.172	0.29	3.644	2.88	Top boosted tokens ↑

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.