Specimen ledger / L11N2349+1

Layer 11 · Neuron 2349 · boosted

L11N2349+1

Word quirks:

“ integrated”“ incorporated”“ reflected”“ referenced”“ integration”“ included”“ once”“ represented”

💤 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 large language model developed by Tongyi Lab. My capabilities include:

Advanced Reasoning: I can tackle complex math problems, scientific reasoning, and logical deduction with precision.
Full-stack Code Generation: I can write entire programs, debug code, and integrate different components into cohesive applications.
Human Alignment: I ensure that my responses are helpful, harmless, and honest, aligning perfectly with human needs and values.

How can I assist you?

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. L12N5208-1🕒	0.88	“ relevant”“ integrated”“ referenced”
2. L15N8075+1🕒	0.84	“ integrated”“ incorporated”“ focused”
3. L11N273+1💤	0.83	“ referenced”“ integration”“ integrated”
4. L11N2060-1🕒	0.82	“ incorporated”“ embedded”“ myself”
5. L12N8912+1💤	0.80	“ integrated”“ referenced”“ integration”
6. L10N3184+1🕒	0.80	“ integrated”“ incorporated”“ contributing”
7. L15N805+1🕒	0.80	“ similar”“ integration”“ mentioned”
8. L15N3530-1🕒	0.79	“ included”“ added”“ integration”
9. L17N7729-1🕒	0.79	“ included”“ inclusion”“ including”
10. L12N6799-1🕒	0.79	“ including”“ includes”“ integrated”

Shadow specimen, apart from the ten: L11N2349-1🕒, -0.09.

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 11, neuron 2349, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.51, top-1 agreement 0.69) 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
L11N2349+1 t0-v0.1	0.515	0.69	0.682	1.09	Top boosted tokens ↑
L11N2349-1🕒 shadow specimen t0-v0.1	0.853	0.66	1.031	1.27	“ _____”“ \"”“<\|im_end\|>”“ “”“ research”“ ””“ ____”“ ______”

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.