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

-0.99

SourceFile

-0.92

FontSize

-0.83

ãĤ¤ãĥĪ

-0.82

BUT

-0.78

CHAT

-0.74

Mech

-0.71

joy

-0.70

 Computing

-0.70

ãĥ¢

-0.70

POSITIVE LOGITS

ched

0.64

 Prediction

0.63

mas

0.62

 Christy

0.61

uties

0.61

aples

0.60

 season

0.60

 Thursday

0.60

 Friday

0.59

 charge

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.