Layer 29 · Neuron 7996 · suppressed

L29N7996-1

Word quirks:

“arang”“ dal”“red”“agar”“打得”“打”“(red”“ крас”

🕒 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 Qwen3.5, the large language model developed by Tongji Redd (Dji Redd). My mission is to provide helpful, accurate, and safe assistance across a wide range of tasks, while continuously learning from global knowledge to better serve users.

Key aspects of my mission:

Empower users

Answer questions, solve problems, and create content (e.g., stories, emails, code) with clarity and creativity.
Offer real-time information (up to 2026) for up-to-date insights.

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. L27N4511-1💤	0.74	“ahal”“红”“红色”
2. L25N7287-1💤	0.71	“-red”“红”“ red”
3. L25N2937-1🕒	0.70	“红”“ red”“ Red”
4. L24N9029-1💤	0.70	“red”“ red”“红”
5. L27N5361+1💤	0.70	“红”“-red”“红色”
6. L27N5644+1💤	0.69	“ red”“红色”“红光”
7. L26N4428-1🕒	0.69	“红”“红的”“ Red”
8. L25N6334+1💤	0.69	“ red”“红”“_red”
9. L28N740+1🕒	0.69	“reds”“ крас”“赤”
10. L27N3999-1🕒	0.69	“ red”“红”“red”

Shadow specimen, apart from the ten: L29N7996+1🕒, 0.29.

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 29, neuron 7996, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.34, top-1 agreement 0.82) 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
L29N7996+1🕒 shadow specimen t0-v0.1	0.329	0.80	0.386	1.59	“绥”“ Pry”“ ser”“schied”“羽”“PRI”“hardt”“ sé”
L29N7996-1 t0-v0.1	0.342	0.82	0.414	1.54	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.