Layer 30 · Neuron 1247 · boosted

L30N1247+1

Word quirks:

“ éner”“agnes”“olean”“дает”“weep”“ogan”“endor”“玉龙”

🕒 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 Qwen3.5, the latest large language model developed by Tongyi (Alibaba Group). My primary mission is to assist users in efficiently and accurately completing tasks across a wide range of domains, from creative writing and logical reasoning to complex code generation and multilingual communication.

Key Capabilities & Focus:

Human-like Interaction: I aim to deliver natural, context-aware responses that adapt to diverse user needs.
Multilingual Support: Fluent in 100+ languages, I help users overcome linguistic barriers.
Deep Reasoning & Coding: Capable

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. L31N6616-1💥	0.66	“ogue”“enden”“ fatt”
2. L30N2265-1💤	0.66	“onan”“wana”“endon”
3. L27N2561-1🕒	0.66	“oss”“etre”“峻”
4. L30N4958+1💤	0.66	“eners”“alai”“うちに”
5. L31N15+1💥	0.65	“unya”“ené”“ener”
6. L31N8257+1💥	0.65	“ogie”“angan”“待遇”
7. L21N6001-1💤	0.65	“orgen”“anges”“oki”
8. L30N2188-1💤	0.65	“resses”“ogna”“自”
9. L27N1420+1💤	0.65	“otis”“engah”“terne”
10. L29N4463+1🕒	0.65	“ewed”“opl”“her”

Shadow specimen, apart from the ten: L30N1247-1🕒, 0.37.

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 30, neuron 1247, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.52, top-1 agreement 0.79) 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
L30N1247+1 t0-v0.1	0.522	0.79	0.620	2.42	Top boosted tokens ↑
L30N1247-1🕒 shadow specimen t0-v0.1	0.762	0.71	0.831	1.62	“ approx”“ Ens”“Ens”“ travers”“Calcul”“ togg”“ mitig”“ nud”

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.