Specimen ledger / L3N2034-1

Layer 3 · Neuron 2034 · suppressed

L3N2034-1

Word quirks:

“ parent”“ subsystem”“ `”“ entity”“ entities”“ `<”“ subclass”“ instantiation”

🕒 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 an AI assistant. My mission is to provide you with the capabilities of a large language model. I can help you with:

Answering questions: I can answer any question you may have about me.
Providing context: I can provide the context for any task you may be performing.
Executing actions: I can execute any action you may require.
Observing results: I can observe the results of any process you may be running.
Generating data: I can generate the data for any system you may be using

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. L14N6674+1🕒	0.65	“ subclass”“ subclasses”“ class”
2. L8N1147-1🕒	0.65	“ `<”“ `”“ subclasses”
3. L7N1755+1🕒	0.64	“ type”“ class”“ `”
4. The False Prophet✅ L3N7155+0.6,L13N4810-0.6	0.63	“ literal”“ severed”“ subsystem”
5. L4N4582+1💥	0.61	“ subclass”“ class”“ subclasses”
6. L6N4622+1🕒	0.61	“ subclass”“ subclasses”“ `”
7. L25N1886-1🕒	0.61	“ entity”“实体”“ entities”
8. L3N2604+1🕒	0.60	“ `”“ instantiated”“ code”
9. L24N3254+1🕒	0.60	“ entities”“ Entity”“ entity”
10. L22N7570+1💤	0.59	“ entity”“ entities”“entity”

Shadow specimen, apart from the ten: L3N2034+1💤, -0.08.

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 3, neuron 2034, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.24, top-1 agreement 0.42) 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
L3N2034+1💤 shadow specimen t0-v0.1	0.764	0.65	0.928	1.47	“⏎”“<\|im_end\|>”“⏎⏎”“⏎ ⏎”“ ⏎”“⏎⏎⏎”“⏎ ⏎”“⏎ ⏎”
L3N2034-1 t0-v0.1	2.242	0.42	2.714	2.41	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.