Layer 18 · Neuron 5992 · boosted

L18N5992+1

Word quirks:

“ suggested”“ suggestion”“ like”“美国人”“ suggestions”“ lives”“ suggests”“ seemed”

💤 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 ChatGPT. My primary goal is to assist users with their daily needs through creative thinking, problem-solving, and content generation. Whether you need help with writing, coding, learning new concepts, or exploring ideas, I aim to provide clear, accurate, and helpful responses tailored to your specific context. Let me know if you have any questions or if you'd like to start a conversation! 😊

*(Note: You may have encountered a slight confusion here—Qwen3.5 is the Chinese model's name, while ChatGPT is the English

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. L13N1005+1🕒	0.84	“ possibly”“ suggested”“ suggesting”
2. L29N3353-1🕒	0.82	“suggest”“ suggested”“ suggest”
3. L17N3555+1💤	0.82	“ suggestions”“ typical”“ analogy”
4. L15N4895-1🕒	0.81	“ suggested”“ suggesting”“ reportedly”
5. L17N6527-1🕒	0.81	“ suggested”“ suggestions”“uggested”
6. L31N1500-1💥	0.80	“ suggest”“建议”“suggest”
7. L15N4483-1💤	0.80	“ similar”“ suggested”“ suggesting”
8. L23N4624+1🕒	0.80	“ suggested”“ suggestion”“建议”
9. L10N6077-1💤	0.79	“ suggested”“ suggestion”“ usually”
10. L16N2755-1🕒	0.79	“ suggestions”“ discussion”“ suggests”

Shadow specimen, apart from the ten: L18N5992-1🕒, 0.03.

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 18, neuron 5992, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.56, top-1 agreement 0.69) 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
L18N5992+1 t0-v0.1	0.564	0.69	0.717	1.23	Top boosted tokens ↑
L18N5992-1🕒 shadow specimen t0-v0.1	1.377	0.55	1.555	1.55	“ responsibility”“ liability”“责任”“免责”“ Liability”“ Responsibility”“承担责任”“責任”

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.