Layer 27 · Neuron 4892 · boosted

L27N4892+1

Word quirks:

“反应”“ reaction”“ Reaction”“reaction”“反應”“Reaction”“ 반응”“ reactions”

🕒 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 Q反应 (Q-Reaction) is a large-scale potential reaction model developed by B2F-REAKP, designed to simulate complex chemical reactions and potential energy surfaces with high accuracy. My core mission is to acceluate the discovery of new materials and reactions by:

Reactions Simulation: Modeling potential reaction pathways and transition reactions with high precision.
Reaction Potential Reaction: Predicting reaction potentials and reaction mechanisms for unknown or poorly understood reactions.
Reaction Discovery: Identifying new reaction pathways and potential reaction products through systematic reaction potential reaction analysis.
**

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. L31N4767+1💥	0.99	“反应”“ reaction”“ Reaction”
2. L25N2396-1🕒	0.99	“反应”“ reaction”“ reactions”
3. L25N6688-1🕒	0.99	“反应”“ reaction”“ Reaction”
4. L26N3552+1🕒	0.98	“ Reaction”“ reaction”“反应”
5. L24N2020-1🕒	0.98	“ reactions”“ reaction”“反应”
6. L22N4781-1🕒	0.98	“反应”“ reaction”“ reactions”
7. L27N4436+1🕒	0.97	“ reaction”“反应”“ Reaction”
8. L27N4485+1🕒	0.97	“ reaction”“ Reaction”“Reaction”
9. Chain Reaction✅ L24N8401-1	0.96	“反应”“ reactions”“ reaction”
10. L20N7250-1🕒	0.96	“反应”“ reaction”“反應”

Shadow specimen, apart from the ten: L27N4892-1💤, 0.14.

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 27, neuron 4892, 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 0.73) 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
L27N4892+1 t0-v0.1	0.693	0.73	0.630	1.45	Top boosted tokens ↑
L27N4892-1💤 shadow specimen t0-v0.1	0.423	0.72	0.548	1.28	“夹”“/exec”“赤”“ardino”“ Estudi”“状态”“ตาล”“执行”

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.