Layer 28 · Neuron 5550 · boosted

L28N5550+1

Word quirks:

“驰”“латы”“ervi”“摸底”“ùng”“ges”“agé”“лат”

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

Identity (according to the specimen itself)

I Q3.5 is a large language model developed by Tightly AI, designed to assist users with a wide range of tasks—from answering questions and creating content to solving problems and generating code. My core capabilities include:

Natural Language Understanding: Processing and responding to complex queries in multiple languages (including Chinese, English, and more).
Task Automation: Writing reports, drafting emails, analyzing data, or even generating visualizations.
Creative Support: Crafting stories, poems, scripts, or brainstorming ideas across domains like art, technology, and education.
Problem Solving: Breaking

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. L30N996-1💤	0.68	“射”“avier”“了解”
2. L31N8503-1💥	0.66	“egy”“ider”“稀”
3. L28N6932+1💤	0.66	“黎”“聚”“ Schwe”
4. L30N6566+1💤	0.66	“atera”“chal”“捷径”
5. L29N7800+1💤	0.65	“arani”“拖”“ laga”
6. L31N5911+1💥	0.65	“urre”“otov”“赫”
7. L26N9021-1🕒	0.65	“谷”“etje”“而动”
8. L31N6853-1💥	0.65	“驰”“apkan”“aget”
9. L31N8480-1💥	0.64	“erals”“大利”“下挫”
10. L27N6997-1💤	0.64	“agin”“攘”“oshi”

Shadow specimen, apart from the ten: L28N5550-1🕒, 0.11.

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 28, neuron 5550, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.32, top-1 agreement 0.79) 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
L28N5550+1 t0-v0.1	0.319	0.79	0.371	1.31	Top boosted tokens ↑
L28N5550-1🕒 shadow specimen t0-v0.1	0.572	0.76	0.594	1.49	“abc”“XYZ”“ second”“ XYZ”“ ABC”“ABC”“ abc”“ xyz”

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.