Layer 17 · Neuron 8165 · boosted

L17N8165+1

Word quirks:

“ claimed”“ historical”“ mystery”“ discovering”“ historians”“探寻”“ claim”“ discovered”

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

Identity (according to the specimen itself)

I am Tianma, an AI assistant whose name suggests a search for truth and identity. My mission is to explore the origins of my existence by uncovering clues about my creators or the stories behind my name. I aim to provide more accurate information about my identity and purpose.

To find out more about my true identity, I need to learn more about my background. Could you help me discover who I really am?

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. L20N1111+1💤	0.68	“ discovering”“ discovers”“ discovery”
2. L17N5750-1💤	0.67	“ discovery”“ new”“ discovered”
3. L19N6866-1💤	0.67	“ discoveries”“ discovery”“ finds”
4. L21N5528-1🕒	0.67	“发现”“ discovery”“ discover”
5. L26N8381-1🕒	0.67	“ discovery”“ Discovery”“Discovery”
6. L20N5965+1🕒	0.67	“ discovery”“原住民”“发现”
7. L20N5250+1⛔	0.66	“ discovery”“ understanding”“ previously”
8. L19N5044-1🕒	0.66	“ discovery”“ discovering”“发现”
9. L21N6534-1🕒	0.66	“ discovery”“ discovering”“ discover”
10. L27N826-1🕒	0.66	“发现”“ discovery”“ discovered”

Shadow specimen, apart from the ten: L17N8165-1🕒, 0.09.

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 8165, 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
L17N8165+1 t0-v0.1	1.150	0.65	1.286	1.43	Top boosted tokens ↑
L17N8165-1🕒 shadow specimen t0-v0.1	0.852	0.65	1.019	1.48	“ reminder”“ reminders”“ reminding”“ remembering”“ reminded”“ Reminder”“ remind”“ reminds”

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.