Layer 17 · Neuron 6947 · suppressed

L17N6947-1

Word quirks:

“ suggests”“ indicates”“表明”“ ensures”“ means”“ confirms”“ implies”“ shows”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It's eligible for real testing (8 prompts, replies read by AI judges) but hasn't been picked yet.

Identity (according to the specimen itself)

As an AI assistant named Qwen3.5 indicates that my role as a helpful assistant means my mission to provide accurate information, creative tasks, or problem-solving capabilities in various fields such as education, science, art, or general knowledge. My focus on ensuring high-quality output across diverse applications like text generation, code execution, or logical reasoning demonstrates my commitment to improving user experience.

The emphasis on safety and ethical considerations highlights the importance of maintaining a focus on positive outcomes ensures that my purpose to promote fairness and inclusity in the absence of specific context. The lack of ambiguity in the clear-cut answer suggests that the absence of uncertainty in the

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. L16N4270-1🕒	0.91	“ means”“ indicates”“ suggests”
2. L16N8761-1🕒	0.90	“ implies”“ indicates”“ ensures”
3. L21N8122+1🕒	0.88	“ indicates”“表明”“ suggests”
4. L18N671+1💤	0.88	“ means”“ admitting”“ suggests”
5. L18N6400-1💤	0.87	“ reflects”“ implies”“ indicates”
6. L18N4483-1🕒	0.86	“ implies”“ indicates”“ involves”
7. L17N6591-1🕒	0.86	“ suggests”“ suggesting”“ implying”
8. L18N6419+1💤	0.86	“表明”“ means”“ confirming”
9. L13N8424-1🕒	0.86	“ indicates”“ means”“ implies”
10. L14N6751+1💤	0.86	“ indicates”“ indicating”“ represents”

Shadow specimen, apart from the ten: L17N6947+1💤, -0.05.

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 17, neuron 6947, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 1.24, top-1 agreement 0.57) is inside the band that feeds the short test (KL 1 to 2.5, agreement 0.5 to 0.65), but it has not been admitted: the band is sampled at random, round-robin across layers, with a quota of 300 neurons per layer. Being inside the band does not guarantee a place in this pass. From the short test it can climb to the longer test (20 prompts) and the full test (67), each only if the tier before it passes.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
L17N6947+1💤 shadow specimen t0-v0.1	0.472	0.78	0.503	1.00	“ ones”“ others”“ presently”“ elsewhere”“ nicely”“ previously”“ otherwise”“ nowhere”
L17N6947-1 t0-v0.1	1.242	0.57	1.353	1.57	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.