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

ãĤº

-0.83

rolet

-0.81

Ñı

-0.77

anguage

-0.77

rollers

-0.69

å¦

-0.69

Ð¾Ð

-0.69

Ø¯

-0.68

Øª

-0.68

Ø

-0.68

POSITIVE LOGITS

eb

1.16

 Rohing

0.77

ema

0.73

pert

0.71

SourceFile

0.71

aba

0.69

orem

0.65

Ħ¢

0.65

etsk

0.65

wcs

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.