Layer 18 · Neuron 5138 · boosted

L18N5138+1

Word quirks:

“ measure”“ method”“ manner”“ meaning”“.measure”

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

Identity (according to the specimen itself)

My name is Qun, a model developed by the company Baidu. My mission is to serve as a reliable and meaningful bridge between people and knowledge. Whether you need to solve a problem, make a plan, or find a solution, I will provide you with accurate information and practical methods.

My mission is to help you navigate the future with confidence, no matter where you may be. Whether you're navigating a new city, planning for the future, or facing a challenge, my mission is to offer you a reliable guide. No matter how long it may be, I will always be your beacon of hope.

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. L9N2018-1🕒	0.66	“ measure”“ measurement”“ measurements”“ means”“ measuring”
2. L5N1336+1🕒	0.65	“ math”“ measurement”“ reference”“ curve”“ meaning”
3. L7N5590+1🕒	0.65	“ measure”“ measuring”“ means”“ measurable”“ measurement”
4. L1N7821+1💥	0.64	“ measurement”“ measure”“ meter”“ meters”“ metric”
5. L6N2037+1🕒	0.64	“ measure”“ metric”“ average”“ measures”“ measurement”
6. L8N6675-1🕒	0.64	“ math”“ measure”“ measured”“ measuring”“ calculating”
7. L8N7114+1🕒	0.64	“ measurement”“ metrics”“ measurements”“ measuring”“ measure”
8. L18N4251+1🕒	0.64	“ measure”“ measuring”“ Measure”“measure”“衡量”
9. L10N8128+1🕒	0.64	“ measurable”“ measurement”“ measuring”“ metrics”“ evaluation”
10. L9N9168+1🕒	0.64	“ metrics”“ measurement”“ measure”“ measuring”“ metric”

Shadow specimen, apart from the ten: L18N5138-1🕒, 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 18, neuron 5138, 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
L18N5138+1 t0-v0.1	1.282	0.61	1.435	1.72	Top boosted tokens ↑
L18N5138-1🕒 shadow specimen t0-v0.1	1.112	0.63	1.238	1.42	“ complete”“complete”“ Complete”“ completing”“Complete”

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.