Layer 25 · Neuron 7758 · suppressed

L25N7758-1

Word quirks:

“根”“ root”“-root”“Root”“root”“ roots”“ Root”“根的”

🕒 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 Roots, a large root model designed by Tongqizhi. My primary mission is to provide accurate, helpful, and trustworthy information across a wide range of topics while adhering to the following core principles:

Accuracy & Reliability: I strive to provide fact-based answers and avoid hallucinations or misinformation.
Safety & Compliance: I strictly adhere to safety guidelines, including the US Roots (United States) regulatory framework, to ensure all content is safe, ethical, and compliant with local laws and regulations.
Helpfulness & Efficiency:

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. L25N5992+1🕒	1.00	“根”“ root”“-root”
2. L25N6980+1🕒	0.99	“根”“ root”“根部”
3. L22N1686-1💤	0.99	“根”“Root”“ roots”
4. L27N4849-1💤	0.99	“根”“-root”“ root”
5. L31N881-1💥	0.99	“根”“ root”“root”
6. L25N5613-1🕒	0.99	“ root”“根”“-root”
7. L24N2893-1🕒	0.99	“ root”“根”“ Root”
8. L24N4438-1🕒	0.99	“根”“ root”“ Root”
9. L25N8650+1🕒	0.98	“根”“ root”“根的”
10. L25N4328-1💥	0.98	“根”“ root”“ roots”

Shadow specimen, apart from the ten: L25N7758+1🕒, 0.29.

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 25, neuron 7758, 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 0.72) 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
L25N7758+1🕒 shadow specimen t0-v0.1	0.641	0.76	0.652	1.37	“ ven”“ veins”“ Ven”“ ax”“ vein”“叶”“ax”“静脉”
L25N7758-1 t0-v0.1	0.695	0.72	0.643	1.23	Top boosted tokens ↑

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.