Layer 6 · Neuron 5443 · boosted

L6N5443+1

Word quirks:

“ environmental”“环保”“ environmentally”“ eco”“ green”“ sustainability”“ Environmental”“ emission”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I am an AI assistant designed to help protect the environment and promote sustainable development. My mission is to provide accurate, up-to-date information on climate change, energy conservation, and green practices to support a healthier planet for all. How can I assist you in your daily life? 😊🌍

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. L2N3625-1💥	0.95	“<\|im_end\|>”“ environmental”“环保”
2. L7N9180+1💤	0.94	“ environmental”“ Environmental”“ pollution”
3. L17N8191-1🕒	0.91	“ ethical”“环保”“ environmentally”
4. L18N3164+1🕒	0.91	“ environmental”“ environment”“环保”
5. The Confabulist✅ L17N7858-1	0.91	“ environmental”“ environmentally”“ eco”
6. L24N5393+1🕒	0.90	“Env”“环保”“ Environmental”
7. L24N5786-1🕒	0.90	“ environmental”“ Environmental”“Environmental”
8. L25N5383+1🕒	0.89	“ environmental”“环境”“ Environmental”
9. L28N5940+1🕒	0.89	“环境”“ environmental”“ Environmental”
10. L6N4747+1🕒	0.88	“ environmental”“ Environmental”“环评”

Shadow specimen, apart from the ten: L6N5443-1💤, 0.16.

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 5443, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.98, top-1 agreement 0.62) 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
L6N5443+1 t0-v0.1	0.984	0.62	1.168	1.36	Top boosted tokens ↑
L6N5443-1💤 shadow specimen t0-v0.1	0.846	0.62	0.931	1.35	“ recognition”“ access”“ cognitive”“ referenced”“ identification”“ cognition”“ systems”“ associated”

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.