Layer 8 · Neuron 1868 · boosted

L8N1868+1

Word quirks:

“⏎”“ ⏎”“ ⏎”“<|im_end|>”“ ‘”“ ______”“ ⏎⏎”“ _____”

🧪 In testing — Answered 0 of 8 prompts so far.

Identity (according to the specimen itself)

I am Qwen3.5, an advanced large language model that integrates knowledge from the field of physics. This allows me to accurately answer questions related to the interaction between high-energy particles and hadrons in high-energy particle collisions.

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. L2N5781-1🕒	0.99	“<\|im_end\|>”“ ⏎”“⏎”
2. L3N3493+1💤	0.99	“⏎”“<\|im_end\|>”“⏎ ⏎”
3. L2N3668+1🕒	0.99	“<\|im_end\|>”“⏎ ⏎”“⏎”
4. L4N6185-1🕒	0.99	“⏎”“ ⏎”“ ‘”
5. L8N5460-1💤	0.98	“⏎”“ ⏎”“ ‘”
6. L8N6692+1💤	0.98	“<\|im_end\|>”“ ⏎”“⏎”
7. L6N4182+1💤	0.98	“⏎”“ ⏎”“ ‘”
8. L7N237-1💤	0.98	“ ‘”“ ⏎⏎”“<\|im_end\|>”
9. L8N1219+1🕒	0.98	“<\|im_end\|>”“ ‘”“ ⏎”
10. L11N2872+1🕒	0.98	“<\|im_end\|>”“ ⏎”“ ⏎⏎”

Shadow specimen, apart from the ten: L8N1868-1💤, -0.12.

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 8, neuron 1868, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. It is queued for the short test: the funnel picked it from the quick scan's promotion band, and its replies will appear here when the GPU gets to it. 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
L8N1868+1 t0-v0.1	1.098	0.59	1.338	1.66	Top boosted tokens ↑
L8N1868-1💤 shadow specimen t0-v0.1	0.495	0.73	0.629	1.08	“ here”“ specific”“ logic”“ usually”“ info”“ actually”“ thinking”“ visually”

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.