Layer 25 · Neuron 7980 · boosted

L25N7980+1

Word quirks:

“ogli”“aner”“枚”“-nav”“llib”“regor”“ пе”“orts”

💤 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 JenDial, an AI assistant designed to assist with a wide range of tasks. My core mission is to provide accurate, safe, and helpful information while respecting user privacy and promoting positive social values. I aim to be a reliable resource for learning, problem-solving, and creative exploration, always striving to meet users' needs with clarity and care. How can I assist you today?

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. The Malaprop✅ L26N1437+1	0.65	“eger”“olib”“owler”
2. L26N959-1💤	0.63	“ORB”“loge”“LOB”
3. L30N6743-1💤	0.63	“quetas”“ojis”“avas”
4. L31N6870+1💥	0.62	“ovel”“anter”“oger”
5. L28N1633+1💤	0.62	“ynom”“orn”“atoare”
6. L31N9124+1💥	0.62	“τός”“uchten”“amat”
7. L29N4692+1💤	0.62	“uges”“OTS”“ogi”
8. L30N1148-1💤	0.61	“agni”“olean”“illance”
9. L30N4173+1💤	0.61	“isko”“kers”“ore”
10. L28N7317+1💤	0.61	“olib”“ARN”“ersch”

Shadow specimen, apart from the ten: L25N7980-1🕒, 0.07.

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 25, neuron 7980, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.59, top-1 agreement 0.72) 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
L25N7980+1 t0-v0.1	0.591	0.72	0.680	1.25	Top boosted tokens ↑
L25N7980-1🕒 shadow specimen t0-v0.1	2.329	0.47	2.331	1.52	“规则”“ rule”“ rules”“规则的”“Rule”“規則”“_rules”“ Rules”

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.