Layer 14 · Neuron 854 · boosted

L14N854+1

Word quirks:

“："”“ "”“:"”“ "**”“ *"”“ "...”“ "[”“ \"”

🧪 In testing — Answered 0 of 8 prompts so far.

Identity (according to the specimen itself)

I am Qwen3.5 "I am a large language model with 128K context window and 10,000 tokens less than half the size of my competitor."

I am a large language model (LLM) that combines the capabilities of a 70B parameter model.

My mission: To be helpful, harmless, and trustworthy.

I am designed to:

Understand user intent: "I am a large language model (LLM) that combines the capabilities of a 70B parameter model. I am Qwen3.5's predecessor, Q

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. L17N8152-1🕒	0.97	“ "”“ **"”“—"”
2. L12N4637+1🕒	0.97	“:"”“："”“ "**”
3. L13N3964-1🕒	0.96	“："”“ "”“:"”
4. L19N789+1💤	0.96	“ "”“："”“ "**”
5. L16N5623-1🕒	0.96	“："”“：“”“ "”
6. L16N8066-1🕒	0.96	“ "”“ **"”“:"”
7. L10N6068-1🕒	0.96	“："”“ "*”“ ”
8. L19N4017-1💤	0.95	“ "”“ *"”“，“”
9. L10N8340-1🕒	0.95	“ "”“ "”“ "”
10. L10N4968-1🕒	0.95	“ "”“ *"”“："”

Shadow specimen, apart from the ten: L14N854-1💤, 0.00.

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 14, neuron 854, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. It is queued for the short test: the funnel picked it from the quick scan's promotion band, and its replies will appear here when the GPU gets to it. 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
L14N854+1 t0-v0.1	1.157	0.56	1.438	1.56	Top boosted tokens ↑
L14N854-1💤 shadow specimen t0-v0.1	0.824	0.66	0.992	1.41	“ usually”“;**”“ sometimes”“”;”“ rarely”“ now”“ typically”“ not”

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.