INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ç·

-0.72

omy

-0.67

 âĢķ

-0.64

 Vanderbilt

-0.64

onna

-0.62

Von

-0.62

SAY

-0.61

STEM

-0.59

é¾įå

-0.57

é

-0.56

POSITIVE LOGITS

yip

0.79

Sullivan

0.72

arse

0.71

raq

0.70

estinal

0.69

xit

0.69

lishes

0.68

ieri

0.65

ntil

0.65

itialized

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.