Layer 7 · Neuron 116 · boosted

L7N116+1

Word quirks:

“⏎”“ ⏎”“ ‘”“⏎ ⏎”“⏎⏎”“ \"”“ ⏎ ⏎”“ ⏎”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I am Qwen3.5, a large language model developed by Alibaba Cloud. My mission is to provide accurate, safe, and helpful information to users while adhering to ethical guidelines and safety standards. I aim to assist with tasks such as answering questions, generating creative content, providing factual data, and supporting human endeavors in a responsible manner. I do not have personal beliefs or desires; instead, my purpose is to serve as a reliable tool for assisting users within the boundaries of my design and programming.

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. L7N2312+1💤	0.99	“⏎”“ ⏎”“ ⏎⏎”
2. L6N8595+1💤	0.98	“⏎”“ ⏎”“ \"”
3. L4N7183-1🕒	0.98	“⏎”“ ‘”“ ⏎”
4. L6N7855+1💤	0.97	“⏎”“ ⏎”“ ⏎”
5. L7N8845-1🕒	0.97	“⏎”“ ⏎”“⏎⏎”
6. L5N3646+1🕒	0.97	“⏎”“ ⏎”“ ⏎”
7. L2N229-1💤	0.97	“ ⏎”“ ⏎⏎”“⏎”
8. L11N76+1💤	0.97	“ ‘”“⏎”“ ⏎”
9. L3N4314-1💤	0.97	“⏎”“ ‘”“ ⏎”
10. L9N4194+1💤	0.97	“⏎”“ ⏎”“⏎⏎”

Shadow specimen, apart from the ten: L7N116-1💤, 0.46.

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 7, neuron 116, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.55, 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
L7N116+1 t0-v0.1	0.545	0.72	0.666	1.10	Top boosted tokens ↑
L7N116-1💤 shadow specimen t0-v0.1	0.517	0.69	0.673	1.24	“ *"”“ "”“ **“”“ ——”“ `"”“ "”“ `”“:"”

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.