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

kees

-0.70

Eat

-0.69

live

-0.65

 Transcript

-0.65

AAF

-0.64

 Subscribe

-0.64

 Disciple

-0.62

 Covenant

-0.62

Ear

-0.62

pmwiki

-0.62

POSITIVE LOGITS

backer

0.76

iband

0.71

plet

0.66

 exha

0.66

 outwe

0.66

ategy

0.66

ndra

0.65

angu

0.64

lest

0.63

aditional

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.