Specimen ledger / L5N6391-1

Layer 5 · Neuron 6391 · suppressed

L5N6391-1

Word quirks:

“ editor”“ Editor”“ Editorial”“ editorial”“ editors”“ edited”“ journalist”“ writer”

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

Identity (according to the specimen itself)

<think>

</think>

By: [Your Name]

---

In the fast-paced world of AI, we explore the latest trends in technology.

---

Introduction

The concept of AI is a testament to technology.

---

Editorial

The idea of AI is a contribution by technology.

---

Review

This article was written by AI.

---

Conclusion

The thoughts of AI are compiled by technology.

---

Acknowledgment

This piece was penned by AI.

The AI judge wasn't sure this holds together.

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. L2N2534+1💥	0.95	“ editor”“ editorial”“ Editorial”
2. L5N4857+1💥	0.94	“ editor”“ editors”“ editorial”
3. L2N2407+1💥	0.92	“ editor”“ Editor”“ editors”
4. L27N731-1🕒	0.90	“ editorial”“ Editor”“ editor”
5. L3N4663+1💥	0.90	“ editor”“ editors”“ Editor”
6. L24N990-1🕒	0.87	“ editors”“ editing”“ editor”
7. L7N316+1🕒	0.86	“ textual”“ corrected”“ editors”
8. L7N8952-1💥	0.85	“ correspondent”“ editorial”“ editor”
9. L6N7543+1🕒	0.84	“ editor”“ news”“ editorial”
10. L8N6056-1🕒	0.84	“ editorial”“ publishing”“ publication”

Shadow specimen, apart from the ten: L5N6391+1🕒, 0.10.

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 5, neuron 6391, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 3.36, top-1 agreement 0.28) 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
L5N6391+1🕒 shadow specimen t0-v0.1	1.343	0.54	1.597	1.63	“ cooking”“ cooked”“ fried”“ frying”“ cook”“ crispy”“ cooks”“ culinary”
L5N6391-1 t0-v0.1	3.362	0.28	3.880	2.76	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.