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

endered

-0.87

ouched

-0.72

aples

-0.71

ushing

-0.69

ingu

-0.68

ockets

-0.67

anned

-0.67

xtap

-0.67

iddled

-0.65

oaded

-0.63

POSITIVE LOGITS

EMA

0.75

BUG

0.71

 STUD

0.68

 Campus

0.66

 certific

0.66

kai

0.62

rum

0.62

tut

0.62

 tutor

0.61

 sorely

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.