Specimen ledger / L4N3704+1

Layer 4 · Neuron 3704 · boosted

L4N3704+1

Word quirks:

“ salary”“ employment”“ compensation”“ employee”“ employees”“ job”“ benefits”“ jobs”

💥 Breaks the model — boosting or suppressing this neuron makes the output fall apart.

Identity (according to the specimen itself)

欢迎加入！🎉

你好！我是新来的员工，正在学习如何快速融入团队。

---

回答：

你好！我是新来的员工，正在学习如何快速融入团队。

---

由于您提供的信息是：

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. L4N7558+1🕒	0.93	“ salary”“ compensation”“ payroll”
2. L8N7129+1🕒	0.93	“ job”“ employee”“ staffing”
3. L8N8065-1🕒	0.93	“ salary”“ employment”“ jobs”
4. L8N6052+1🕒	0.92	“ job”“ jobs”“ employee”
5. L9N7722+1🕒	0.90	“ job”“ employment”“ jobs”
6. L8N3058+1🕒	0.89	“ employment”“ jobs”“ workforce”
7. L6N8109-1🕒	0.89	“ job”“ employment”“ jobs”
8. L7N6030-1🕒	0.88	“ employment”“ Employment”“ workers”
9. L19N6263-1💤	0.88	“ employment”“ Employment”“ employer”
10. L7N6358-1🕒	0.88	“ job”“ jobs”“ employment”

Shadow specimen, apart from the ten: L4N3704-1🕒, -0.16.

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 4, neuron 3704, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.81, top-1 agreement 0.37) 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
L4N3704+1 t0-v0.1	2.813	0.37	3.246	2.48	Top boosted tokens ↑
L4N3704-1🕒 shadow specimen t0-v0.1	1.196	0.55	1.413	1.72	“ literal”“ metaphor”“ literally”“ words”“ —”“ word”“ *"”“ ——”

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.