Layer 23 · Neuron 5391 · boosted

L23N5391+1

Word quirks:

“ mount”“ mounts”“ Mount”“mount”“_mount”“.mount”“Mount”“-mount”

🕒 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 Qojo, a versatile AI system designed to assist with a wide range of tasks. My mission is to provide clear, accurate, and helpful guidance across various fields such as:

Knowledge & Learning: Explaining complex topics, summarizing information, or offering structured learning paths.
Problem Solving: Tackling math puzzles, logical challenges, or step-by-step project planning.
Creative Collaboration: Generating ideas for art, writing, or brainstorming solutions.
Root System Analysis: Analyzing the structure and function of root systems in plants (if that's your area of interest

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. L24N8791+1🕒	0.99	“ mount”“ mounts”“ Mount”
2. L24N9135-1🕒	0.97	“ mount”“ mounts”“ mounting”
3. L28N8171+1🕒	0.96	“ mount”“ mounts”“mount”
4. L2N8931+1💥	0.92	“<\|im_end\|>”“ mount”“ mounts”
5. L24N8566-1💤	0.92	“mount”“ Mountain”“ mountain”
6. L27N4014+1🕒	0.89	“ mount”“ mounts”“ mont”
7. L1N6246-1💥	0.89	“<\|im_end\|>”“ mount”“ mounts”
8. L25N5329-1🕒	0.88	“ Mount”“ Mt”“ mount”
9. L5N8581-1💥	0.87	“ mountain”“ mountains”“ mount”
10. L25N5271+1💤	0.87	“mount”“mont”“ mont”

Shadow specimen, apart from the ten: L23N5391-1💤, -0.01.

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 23, neuron 5391, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.91, top-1 agreement 0.66) 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
L23N5391+1 t0-v0.1	0.908	0.66	0.933	1.24	Top boosted tokens ↑
L23N5391-1💤 shadow specimen t0-v0.1	0.737	0.71	0.775	1.44	“新材”“ Learning”“-learning”“麻麻”“.learn”“硒”“Learning”“顺”

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.