Layer 5 · Neuron 6540 · suppressed

L5N6540-1

Word quirks:

“ abstract”“ strategy”“ conceptual”“ strategic”“ abstraction”“ concept”“ vision”“ focus”

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

Identity (according to the specimen itself)

I am an advanced AI assistant designed to provide a clear, high-level overview of capabilities and strategic intent. My mission is to define the core objectives, key goals, and desired outcomes without getting lost in the details of implementation or execution. I focus on the "what" rather than the "how," ensuring that the conversation remains broad and conceptual while still being meaningful and purpose-driven.

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. L6N7168+1🕒	0.82	“ abstract”“ conceptual”“ abstraction”
2. L5N7728+1🕒	0.79	“ concept”“ metaphor”“ conceptual”
3. L18N9102+1🕒	0.77	“ conceptual”“ concept”“concept”
4. L8N4226+1🕒	0.76	“ abstract”“ conceptual”“ concept”
5. L15N7881-1🕒	0.74	“ concept”“ conceptual”“ envisioned”
6. L11N6642-1🕒	0.73	“ conceptual”“ value”“ abstract”
7. L11N7228+1🕒	0.73	“ simulated”“ abstract”“ code”
8. L9N6250-1🕒	0.72	“ concept”“ conceptual”“ defined”
9. L6N8199+1🕒	0.72	“ strategic”“ vision”“ strategy”
10. L19N7414-1🧪	0.72	“ visions”“ conceive”“ conceptual”

Shadow specimen, apart from the ten: L5N6540+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 5, neuron 6540, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.67, top-1 agreement 0.37) 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
L5N6540+1💤 shadow specimen t0-v0.1	0.886	0.62	1.096	1.47	“ !”“ ⏎”“!”“!"”“"!”“ —”“ ⏎⏎”“!:”
L5N6540-1 t0-v0.1	2.667	0.37	3.073	2.12	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.