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

xes

-0.87

govtrack

-0.83

enge

-0.79

ranch

-0.75

oras

-0.74

cue

-0.71

cÃ©

-0.70

ilon

-0.70

uts

-0.70

options

-0.69

POSITIVE LOGITS

 gloss

0.67

 persist

0.66

LTD

0.66

USS

0.64

NS

0.63

HMS

0.61

 treated

0.59

 constit

0.59

 nails

0.59

 subs

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.