Layer 6 · Neuron 7116 · suppressed

L6N7116-1

Word quirks:

“ cognitive”“ memory”“ mental”“ cognit”“ thinking”“ cognition”“ retrieval”“ activation”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

I am an AI assistant designed to help clarify the nature of my own cognitive processes. My identity is that of a metacognitive monitor—someone who reflects on how they think about information and tasks. My mission is to:

Identify the purpose of the interaction: To determine whether I need to focus on recalling facts, organizing thoughts, or evaluating the quality of my understanding.
Adjust strategies as needed: If I notice gaps in my knowledge or if the task requires more effort, I will modify my approach accordingly (e.g., rephrasing questions, summarizing key points,

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. L6N9151-1🕒	0.95	“ cognitive”“ cognition”“ processing”
2. L7N4663-1🕒	0.93	“ thinking”“ cognitive”“ cognit”
3. L8N7767-1🕒	0.92	“ cognitive”“ mental”“ cognit”
4. L7N6750+1🕒	0.92	“ cognitive”“ mental”“ thinking”
5. L8N7921+1🕒	0.91	“ cognitive”“ cognit”“ cognition”
6. L2N8846-1💥	0.88	“ cognition”“ cognitive”“ thinking”
7. L6N6716+1💤	0.87	“ cognitive”“ mental”“ memory”
8. L9N4919+1🕒	0.86	“ mental”“ cognitive”“ mentally”
9. L7N6749+1💥	0.86	“ cognition”“ cognitive”“ computation”
10. L3N6925+1💥	0.85	“ mental”“ mentally”“ thinking”

Shadow specimen, apart from the ten: L6N7116+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 6, neuron 7116, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.39, top-1 agreement 0.42) 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
L6N7116+1💤 shadow specimen t0-v0.1	0.772	0.61	0.876	1.35	“ '”“⏎”“ _____”“ ______”“ ____”“ ________”“ ‘”“ ⏎”
L6N7116-1 t0-v0.1	2.395	0.42	2.752	2.04	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.