INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

Iv

-0.72

bread

-0.69

wives

-0.68

tein

-0.68

earchers

-0.66

ombo

-0.64

Avg

-0.63

ebook

-0.62

 Scrolls

-0.62

EO

-0.61

POSITIVE LOGITS

 Brow

0.64

istor

0.60

frequency

0.60

itent

0.59

 Hicks

0.58

 Mercer

0.58

canon

0.56

 tampering

0.56

 Berman

0.55

ihar

0.55

Activations Density 0.000%

No Known Activations

This feature has no known activations.