INDEX

Explanations

Code/abbreviations/names

np_max-act · gemini-2.0-flash

the neuron detects capitalized headings, short uppercase acronyms, and other prominent named-entity tokens (e.g., CT, PWM, QR, NS, proper names).

oai_token-act-pair · gpt-5-mini Triggered by @vetterc0

New Auto-Interp

Configuration

andyrdt/saes-qwen2.5-7b-instruct/resid_post_layer_15/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.15.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

🌸

-0.08

Rak

-0.08

 effet

-0.07

 Padres

-0.07

 maison

-0.07

 Marriott

-0.07

 Citadel

-0.06

.extra

-0.06

akah

-0.06

ティー

-0.06

POSITIVE LOGITS

CREMENT

0.07

 Twenty

0.07

nearest

0.07

觳

0.07

⟵

0.06

 currency

0.06

购票

0.06

⚁

0.06

_FAR

0.06

 represented

0.06

Activations Density 0.690%

Code/abbreviations/names

the neuron detects capitalized headings, short uppercase acronyms, and other prominent named-entity tokens (e.g., CT, PWM, QR, NS, proper names).

No Comments

No Known Activations