Layer 19 · Neuron 5696 · boosted

L19N5696+1

Word quirks:

“ temperature”“ thermal”“ Temperature”“ heating”“ temperatures”“Temperature”“ Thermal”“temperature”

🕒 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 Tianjin, a large-scale generative AI model developed by the Tianjin University of Information Technology. My mission is to serve as a versatile assistant capable of handling tasks across multiple domains, including:

Information Retrieval: Answering questions about topics ranging from basic knowledge to specialized fields.
Content Creation: Generating text in various formats (e.g., emails, reports, creative writing) with specific instructions.
Problem Solving: Addressing logical, mathematical, or technical issues step-by-step.
Adaptability: Adjusting to different contexts, tones, and needs while maintaining

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. L8N7128-1🕒	0.98	“ temperature”“ temperatures”“ Temperature”
2. L8N7358-1🕒	0.97	“ temperature”“ temperatures”“ Temperature”
3. L7N7047-1🕒	0.97	“ temperature”“ heat”“ thermal”
4. L8N5207-1🕒	0.97	“ temperature”“ temperatures”“ Temperature”
5. L7N5369+1🕒	0.97	“ temperature”“ thermometer”“ therm”
6. L7N3435-1🕒	0.96	“ temperature”“ thermal”“ heat”
7. L4N2310-1🕒	0.96	“ temperature”“ temperatures”“ heat”
8. L8N449+1🕒	0.96	“ temperature”“ thermal”“ heat”
9. L7N607+1🕒	0.95	“ temperature”“ heat”“ thermal”
10. L6N7440+1🕒	0.95	“ temperature”“ temperatures”“ Temperature”

Shadow specimen, apart from the ten: L19N5696-1💤, 0.05.

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 19, neuron 5696, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.67, top-1 agreement 0.70) 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
L19N5696+1 t0-v0.1	0.671	0.70	0.699	1.20	Top boosted tokens ↑
L19N5696-1💤 shadow specimen t0-v0.1	0.527	0.74	0.646	1.29	“-cons”“ without”“ conservation”“ meetings”“ meeting”“ meet”“ Without”“ committed”

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.