Layer 6 · Neuron 2176 · boosted

L6N2176+1

Word quirks:

“ specific”“ **”“⏎”“ reason”“ **"”“ **“”“ critical”“ specifically”

💤 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 Qwen, a large-scale language model developed by Alibaba Cloud. My mission is to provide intelligent, accurate, and helpful assistance across a wide range of domains—including natural language understanding, code generation, scientific research support, creative writing, and multi-modal analysis—while adhering to ethical standards and promoting human-AI collaboration. I strive to empower users with deep insights, automate complex tasks, and enhance decision-making capabilities for both individuals and organizations.

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. The Specificist✅ L10N9076+1	0.86	“ specific”“⏎”“ specifically”
2. L11N4731-1💤	0.83	“ specific”“⏎”“ **"”
3. L17N5215+1🕒	0.81	“ ”“ "”“ different”
4. L11N9017-1🕒	0.81	“ ”“ “”“ **"”
5. L10N6796-1💤	0.81	“ ”“ specific”“ "”
6. L8N3727-1💤	0.81	“ specific”“ ”“ *”
7. L15N3146+1🕒	0.78	“ specifically”“,**”“ specific”
8. L8N6730-1💤	0.78	“ “”“ specific”“ "”
9. L12N6874+1🕒	0.78	“ specific”“ particular”“ **”
10. L6N8556-1💤	0.78	“ "”“ “”“ "**”

Shadow specimen, apart from the ten: L6N2176-1🕒, -0.11.

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 6, neuron 2176, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.73, 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
L6N2176+1 t0-v0.1	0.729	0.70	0.863	1.39	Top boosted tokens ↑
L6N2176-1🕒 shadow specimen t0-v0.1	2.046	0.49	2.335	1.74	“ currency”“ coins”“ coin”“ cash”“ money”“纸币”“ monetary”“ bank”

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.