Layer 28 · Neuron 4753 · suppressed

L28N4753-1

Word quirks:

“ lib”“助”“li”“ Lib”“司”“ li”“⇥lib”“lib”

💤 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, the latest language model in the Qwen series developed by Tonglihua (Tongliblihua). My primary mission is to provide helpful, accurate, and safe assistance across a wide range of tasks, from answering questions and solving problems to creative writing, code generation, and complex reasoning.

Key Features & Capabilities:

Largely Language Understanding: I support over 100 languages globally, ensuring cross-lingual communication and localization.
Advanced Reasoning: I excel in **mathematics, natural sciences

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. L29N5743+1🕒	0.88	“ lib”“ Lib”“lib”
2. L26N5858+1🕒	0.86	“ Lib”“-lib”“ librarian”
3. L30N282+1💥	0.85	“lib”“ lib”“Lib”
4. L28N6433-1💤	0.84	“ lib”“lib”“ Lib”
5. L27N4236-1💤	0.83	“lib”“ lib”“ lab”
6. L26N2642-1🕒	0.83	“ lib”“ Lib”“-lib”
7. L27N5740+1💤	0.83	“Lib”“ lib”“lib”
8. L29N4523+1🕒	0.82	“ lib”“lib”“Lib”
9. L28N5037-1🕒	0.81	“lib”“ lib”“Lib”
10. L26N2300-1🕒	0.81	“库”“图书馆”“ library”

Shadow specimen, apart from the ten: L28N4753+1🕒, 0.14.

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 28, neuron 4753, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.33, top-1 agreement 0.83) 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
L28N4753+1🕒 shadow specimen t0-v0.1	0.676	0.72	0.645	1.45	“ Project”“ project”“ spec”“project”“项目”“_spec”“Project”“-spec”
L28N4753-1 t0-v0.1	0.325	0.83	0.433	1.21	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.