Layer 20 · Neuron 2419 · boosted

L20N2419+1

Word quirks:

“ strings”“ string”“ String”“strings”“ expressions”“ Strings”“ represented”“ representing”

🕒 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'm a large language model, an AI system designed to assist with information, creativity, problem-solving, and many other tasks. My mission is to empower users by providing clear, accurate, and helpful responses—whether that's explaining complex ideas, creating content, coding, or even just sharing a smile through a joke. I strive for accuracy, empathy, and adaptability, always aiming to make interactions meaningful and positive. How can I help 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. L24N832+1🕒	0.91	“ strings”“ string”“ String”
2. L22N6049+1🕒	0.90	“ strings”“ Strings”“字符串”
3. L22N3269-1💤	0.89	“ strings”“ string”“字符串”
4. L31N8212+1💥	0.88	“string”“ string”“ strings”
5. L0N5561+1💥	0.87	“string”“String”“-string”
6. L27N2365+1🕒	0.87	“ string”“ strings”“-string”
7. L24N2930-1🕒	0.86	“ string”“ strings”“-string”
8. L2N7373-1💥	0.85	“弦”“ string”“ strings”
9. L2N2493-1💥	0.85	“ string”“<\|im_end\|>”“-string”
10. L23N5659-1🕒	0.84	“弦”“ strings”“strings”

Shadow specimen, apart from the ten: L20N2419-1🕒, -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 20, neuron 2419, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.69, top-1 agreement 0.71) 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
L20N2419+1 t0-v0.1	0.689	0.71	0.843	1.33	Top boosted tokens ↑
L20N2419-1🕒 shadow specimen t0-v0.1	0.684	0.70	0.782	1.43	“刚从”“ captured”“ffen”“=="”“ hired”“一新的”“ khỏi”“ remembered”

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.