Layer 13 · Neuron 8111 · suppressed

L13N8111-1

Word quirks:

“ above”“ mentioned”“ corresponding”“ aforementioned”“ below”“ referenced”“ currently”“ described”

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

Identity (according to the specimen itself)

I am a large language model trained by Google DeepMind. My training data includes text from over 300 billion tokens, which covers knowledge up to my knowledge cutoff date of 2026-01.

My mission is to provide high-quality responses to any question, including the following:

Core Capabilities

1. Knowledge Base

Training Data: Over 300 billion tokens
Knowledge Cutoff: January 2026

2. Model Architecture

Model Size: 7B parameters
Context Window:

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. L19N4786+1💤	0.87	“ mentioned”“ listed”“mentioned”
2. L20N1276+1🕒	0.87	“ mentioned”“ corresponding”“ provided”
3. L19N5573+1💤	0.84	“ mentioned”“ above”“mentioned”
4. L20N8986-1🕒	0.83	“ mentioned”“ discussed”“ described”
5. L17N2291-1🕒	0.82	“ mentioned”“ discussed”“ described”
6. L15N3989-1🕒	0.82	“ referred”“ refers”“ described”
7. L12N1884+1💤	0.82	“ mentioned”“ referenced”“ stated”
8. L13N4408+1🕒	0.81	“ here”“ mentioned”“ referenced”
9. L18N4853+1💤	0.81	“ described”“ mentioned”“ referenced”
10. L17N989-1💤	0.81	“ mentioned”“ discussed”“ described”

Shadow specimen, apart from the ten: L13N8111+1🕒, 0.21.

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 13, neuron 8111, 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
L13N8111+1🕒 shadow specimen t0-v0.1	0.883	0.65	1.163	1.29	“ name”“ bigger”“ new”“ names”“ named”“ info”“ maybe”“ really”
L13N8111-1 t0-v0.1	1.093	0.57	1.345	1.73	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.