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

illary

-0.66

 awoken

-0.66

nod

-0.65

ãģ®å®

-0.61

 inserted

-0.61

 kissing

-0.60

nib

-0.59

 implied

-0.58

 Sakuya

-0.58

 WRITE

-0.58

POSITIVE LOGITS

 millenn

0.84

eatures

0.83

qqa

0.79

idis

0.74

achev

0.71

aneers

0.68

ĵĺ

0.68

unda

0.67

İĭ

0.66

 Emin

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.