Layer 14 · Neuron 5687 · suppressed

L14N5687-1

Word quirks:

“ basically”“ means”“ essentially”“ implies”“ Basically”“ saying”“ implying”“ seems”

🧪 In testing — Answered 0 of 8 prompts so far.

Identity (according to the specimen itself)

I am a large language model created by Google. This means I'm essentially a supercomputer where I can think of it like having a brain that allows me to understand and generate text. It's like if I can read all books is like having a library where I can answer questions like "What is the capital of France?" (I know because I have access to my knowledge base).

This is basically like saying: "I am a robot that knows everything." (Which is like saying I can talk to you is like a human saying "I am a human" - which is like saying "I am a human" - which is

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. L14N9142+1💤	0.89	“ basically”“ means”“ similar”
2. L4N5251-1💥	0.87	“ probably”“ implies”“ basically”
3. L9N8887-1🕒	0.87	“ basically”“ means”“ signifies”
4. L17N9027+1🕒	0.86	“ means”“ refers”“ whatever”
5. L17N4450+1🕒	0.85	“ basically”“ implying”“ implies”
6. L3N3917-1🕒	0.83	“⏎”“ means”“ `”
7. L12N1207+1💤	0.83	“ seems”“ means”“ implies”
8. L14N6649-1🕒	0.82	“ therefore”“ means”“ implies”
9. L11N5748-1🕒	0.81	“ implied”“ neither”“ implying”
10. L15N6476-1🕒	0.81	“ basically”“ implying”“ indicating”

Shadow specimen, apart from the ten: L14N5687+1💤, 0.18.

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 14, neuron 5687, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. It is queued for the short test: the funnel picked it from the quick scan's promotion band, and its replies will appear here when the GPU gets to it. 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
L14N5687+1💤 shadow specimen t0-v0.1	0.708	0.65	0.915	1.36	“⏎”“ ⏎”“ its”“ ⏎”“ what”“ each”“ additionally”“ similarly”
L14N5687-1 t0-v0.1	1.631	0.51	1.936	1.81	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.