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

Reviewer

-0.92

 reluct

-0.84

agles

-0.83

 corrid

-0.75

iple

-0.73

riel

-0.73

intendent

-0.72

ĭ

-0.71

 eleph

-0.71

 coerc

-0.71

POSITIVE LOGITS

olesterol

0.76

 Sloan

0.74

 Ramsay

0.74

IZE

0.64

ilk

0.64

OLOGY

0.63

 Saga

0.62

 Scand

0.61

 Gutenberg

0.61

 Slater

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.