Layer 27 · Neuron 6713 · suppressed

L27N6713-1

Word quirks:

“ lib”“繪”“lib”“原”“orp”“⇥lib”“isses”“程”

💤 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 Q7, a large-scale language model developed by the University of Cambridge as part of the Cambridge NLP research program. My primary purpose is to serve as a versatile AI assistant capable of understanding and generating human language with a strong emphasis on accuracy, clarity, and ethical alignment.

My Mission

My mission is to empower users by:

Answering questions with precision, from complex technical concepts to everyday queries.
Solving problems through logical reasoning, code generation, or creative problem-solving.
Facilitating learning by explaining

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🕒	0.83	“ lib”“ Lib”“lib”
2. L28N4753-1💤	0.81	“ lib”“助”“li”
3. L26N5858+1🕒	0.78	“ Lib”“-lib”“ librarian”
4. L26N1429-1💤	0.76	“llib”“角”“-lib”
5. L26N2300-1🕒	0.71	“库”“图书馆”“ library”
6. L28N6433-1💤	0.70	“ lib”“lib”“ Lib”
7. L27N4236-1💤	0.70	“lib”“ lib”“ lab”
8. L27N5740+1💤	0.69	“Lib”“ lib”“lib”
9. L29N6926+1🕒	0.68	“ liberal”“立”“ Liberal”
10. L23N1138+1🕒	0.68	“ library”“-library”“ religion”

Shadow specimen, apart from the ten: L27N6713+1💤, 0.42.

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 27, neuron 6713, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.43, top-1 agreement 0.79) 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
L27N6713+1💤 shadow specimen t0-v0.1	0.485	0.81	0.489	1.46	“参数”“记”“记录”“蒙”“说明”“ede”“卷”“减”
L27N6713-1 t0-v0.1	0.426	0.79	0.444	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.