Layer 17 · Neuron 2269 · suppressed

L17N2269-1

Word quirks:

“ evaluations”“ evaluation”“ checks”“ evalu”“ assessment”“ visibly”“ nod”“ assessments”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

I am Qianwen3.5, a large-scale AI model developed by Tongyi Lab. My primary mission is to perform high-quality tasks across various domains, including natural language processing, computer vision, and code generation. I aim to provide reliable, efficient, and secure outputs for users who may use me for research, development, or general-purpose applications.

To ensure this, I prioritize:

Accuracy: Delivering precise results in specialized fields like scientific computing or legal documents.
Security: Avoiding sensitive data exposure through strict validation protocols.
Efficiency: Reducing latency in critical systems via

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. L9N8190+1🕒	0.94	“ assessment”“ evaluation”“ assessing”
2. L10N8419+1🕒	0.93	“ evaluations”“ evaluation”“ assessments”
3. L7N2482-1🕒	0.92	“ assessments”“ assessment”“ evidence”
4. L11N3123+1🕒	0.91	“ report”“ presently”“ assessed”
5. L2N9057+1💥	0.91	“ assessment”“<\|im_end\|>”“ evaluation”
6. L5N6583+1🕒	0.91	“ assessment”“ evaluation”“ assessing”
7. L16N8799+1🕒	0.90	“ evaluation”“ assessment”“ evaluations”
8. L5N6448-1💥	0.90	“ evaluation”“ evaluations”“ assessment”
9. L14N7024-1🕒	0.90	“ assessments”“ assessment”“ evaluations”
10. L11N2608-1💤	0.90	“ assessment”“ evaluation”“ based”

Shadow specimen, apart from the ten: L17N2269+1🕒, 0.04.

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 17, neuron 2269, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.77, top-1 agreement 0.70) 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
L17N2269+1🕒 shadow specimen t0-v0.1	1.511	0.54	1.797	1.42	“ tale”“ story”“ tales”“ stories”“ narrative”“ storytelling”“ storyt”“ heroes”
L17N2269-1 t0-v0.1	0.767	0.70	0.837	1.15	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.