Layer 24 · Neuron 5783 · suppressed

L24N5783-1

Word quirks:

“ hair”“ Hair”“hair”“Hair”“ hairs”“头发”“发型”“ shorter”

💤 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 QSL, a specialized AI model developed by the QSL Team (a team focused on advanced system architecture and dialogue systems). My primary mission is to serve as a high-fidelity digital partner for users who need precise, structured, and reliable information retrieval or system interaction.

Unlike general-purpose models that prioritize broad conversation or creative generation, my focus is on:

Systemic Precision: Providing answers that align with specific technical schemas or predefined data structures.
Role-Driven Interaction: Acting as a functional agent within a larger system, handling tasks like data parsing, ticket management,

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. L2N5249-1💥	0.97	“ hair”“ hairs”“<\|im_end\|>”
2. L31N4437-1💥	0.97	“hair”“ Hair”“ hair”
3. L18N2615+1🕒	0.95	“ hair”“短发”“ hairstyle”
4. L26N3888+1🕒	0.94	“ hair”“ Hair”“hair”
5. L7N6680+1🕒	0.94	“ hair”“ Hair”“ hairstyle”
6. L17N8911-1🕒	0.94	“ Hair”“ hair”“Hair”
7. The Wig Fitter✅ L5N7266+1	0.92	“ hair”“ hairstyle”“ Hair”
8. L19N5750-1🕒	0.92	“ hair”“ shampoo”“ haircut”
9. L6N7506+1🕒	0.90	“ hair”“ Hair”“ salon”
10. L21N8046+1🕒	0.89	“ hair”“ growth”“ unk”

Shadow specimen, apart from the ten: L24N5783+1🕒, 0.16.

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 24, neuron 5783, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.74, 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
L24N5783+1🕒 shadow specimen t0-v0.1	1.260	0.65	1.300	1.45	“ materials”“ material”“材料”“ Materials”“ Material”“-material”“材料的”“material”
L24N5783-1 t0-v0.1	0.741	0.73	0.762	1.27	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.