INDEX

Explanations

code and file paths

np_max-act · gemini-2.0-flash

The neuron reliably lights up on short uppercase alphabetic tokens—particularly two‐letter codes or abbreviations—regardless of surrounding syntax.

oai_token-act-pair · o4-mini Triggered by @xinyanhu8

New Auto-Interp

Configuration

andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.11.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

_Collections

-0.07

 لا

-0.06

평

-0.06

едж

-0.06

When

-0.06

Appending

-0.06

.faces

-0.06

 συγκ

-0.06

Dar

-0.06

 воз

-0.06

POSITIVE LOGITS

 been

0.08

 translates

0.07

erase

0.07

 Composer

0.06

populate

0.06

NavigationView

0.06

hop

0.06

 annotation

0.06

 dream

0.06

 policies

0.06

Activations Density 0.165%

code and file paths

The neuron reliably lights up on short uppercase alphabetic tokens—particularly two‐letter codes or abbreviations—regardless of surrounding syntax.

No Comments

No Known Activations