Layer 23 · Neuron 8173 · suppressed

L23N8173-1

Word quirks:

“ dependence”“dependence”“依赖性”“依赖”“ Depend”“ dependency”“ dependent”“-depend”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I am Tianqi (天启), a large language model developed by CLOTT (China Online Learning Technology Group). My mission is to provide users with high-quality, accurate, and helpful information across a wide range of topics—whether it's answering questions, writing creative content, analyzing data, or assisting in decision-making. I aim to be a reliable tool for learning, creativity, and problem-solving while adhering to safety guidelines and ethical standards. How can I assist you today? 🌟

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. L20N4301-1🕒	1.00	“ dependence”“ Depend”“依赖”
2. L25N1924-1🕒	0.99	“ Depend”“ dependence”“依赖”
3. L23N4095-1🕒	0.99	“依赖”“ dependence”“ dependent”
4. L25N936-1🕒	0.99	“依赖”“ dependent”“ dependence”
5. L19N5311+1🕒	0.98	“ dependency”“ dependence”“依赖”
6. L31N1030+1💥	0.98	“依赖”“ depend”“ dependence”
7. L21N8781-1💤	0.98	“依赖”“ dependency”“ dependence”
8. L22N5949-1💤	0.97	“依赖于”“依赖”“-depend”
9. L24N5943+1🕒	0.97	“依赖”“ dependency”“ dependence”
10. L31N7335+1💥	0.96	“ dependence”“ dependency”“依赖”

Shadow specimen, apart from the ten: L23N8173+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 23, neuron 8173, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.65, top-1 agreement 0.76) 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
L23N8173+1💤 shadow specimen t0-v0.1	0.683	0.73	0.709	1.36	“五指”“FL”“/fl”“ FL”“entai”“ Khu”“хами”“erval”
L23N8173-1 t0-v0.1	0.649	0.76	0.703	1.15	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.