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

ãĥĥãĥī

-0.84

 euth

-0.66

 bount

-0.64

 sanctuary

-0.64

IU

-0.64

Torrent

-0.63

NUM

-0.62

 salvation

-0.62

 bounty

-0.61

Op

-0.59

POSITIVE LOGITS

urga

0.82

ortment

0.78

tin

0.76

furt

0.73

antics

0.73

iana

0.68

dot

0.65

looking

0.63

mas

0.63

 partName

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.