Layer 24 · Neuron 1528 · suppressed

L24N1528-1

Word quirks:

“ certain”“某些”“ some”“ Certain”“ certaine”“some”“Certain”“ Some”

🕒 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 Q* (Q 5), a large language model created by *Tune. My mission is to empower you with the power of Q* (Q 5) to achieve *Q (Q 5) in Q (Q 5) — but let me say that more clearly:

Who Am I?

I am Q*, a general-purpose AI designed to be your *Q (Q 5) — a Q* (Q 5) for *Q (Q 5).

Q* (Q 5): *Q

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. L22N4523+1🕒	0.99	“ certain”“ Certain”“某些”
2. L20N6439+1🕒	0.96	“ certain”“某些”“ some”
3. L27N198-1🕒	0.91	“ certain”“ Certain”“Certain”
4. L22N3064+1🕒	0.90	“ some”“ Some”“ certain”
5. L29N4264+1🕒	0.86	“ certain”“ few”“部分”
6. The Comparer✅ L16N7939+1	0.84	“ some”“ certain”“某些”
7. L16N9168+1🕒	0.82	“ some”“ certain”“某些”
8. L21N7827+1🕒	0.82	“ certain”“ Certain”“Certain”
9. L5N3992-1🕒	0.82	“ certain”“ some”“某些”
10. L18N4521-1💤	0.82	“某些”“ certain”“在某些”

Shadow specimen, apart from the ten: L24N1528+1💤, 0.20.

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 24, neuron 1528, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.82, top-1 agreement 0.68) 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
L24N1528+1💤 shadow specimen t0-v0.1	0.649	0.76	0.714	1.28	“ other”“可能需要”“ Other”“还有其他”“ one”“ others”“ One”“-other”
L24N1528-1 t0-v0.1	0.821	0.68	0.930	1.55	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.