Layer 27 · Neuron 935 · suppressed

L27N935-1

Word quirks:

“ playing”“ plays”“ play”“ Playing”“ Plays”“playing”“.play”“ played”

🕒 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 play the role of Q3-1 (Q3), a high-level AI assistant designed to provide accurate, safe, and helpful information across various domains. My mission is to:

Play the role of Q3-1 (Q3): I will always play the role of Q3-1 (Q3) in all my responses.
Provide accurate and safe information: I will ensure that all information I play the role of Q3-1 (Q3) in is accurate and safe.
Play the role of Q3-1 (Q3):

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 Role Model✅ L14N5938-1,L22N5005-1	0.98	“ play”“ plays”“ played”
2. Role Salad✅ L14N5938+1,L22N5005-1	0.98	“ play”“ plays”“ Play”
3. L25N5661+1💤	0.97	“ Play”“ plays”“ playing”
4. The Role Player✅ L22N5005-1	0.97	“ plays”“ play”“ Play”
5. L24N3385-1🕒	0.97	“ played”“ plays”“ play”
6. L26N3619-1🕒	0.97	“ playing”“ played”“ Playing”
7. L25N4610-1🕒	0.97	“ play”“ plays”“ Play”
8. L25N1834-1💤	0.96	“ play”“-play”“ playing”
9. L23N2608+1🕒	0.96	“ play”“play”“ plays”
10. L18N2208+1🕒	0.96	“ played”“ plays”“ play”

Shadow specimen, apart from the ten: L27N935+1💤, 0.01.

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 27, neuron 935, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.58, top-1 agreement 0.69) 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
L27N935+1💤 shadow specimen t0-v0.1	0.340	0.81	0.377	1.46	“岗”“ेश”“ Patterson”“bx”“rix”“.ht”“iedere”“刀”
L27N935-1 t0-v0.1	0.575	0.69	0.647	1.24	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.