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

SAP

-0.76

 Host

-0.73

TBA

-0.68

 Brand

-0.66

 Creatures

-0.65

 Suggest

-0.63

)</

-0.63

ventory

-0.61

 Ended

-0.60

rab

-0.60

POSITIVE LOGITS

alus

0.78

naissance

0.66

aunder

0.66

roller

0.65

rawn

0.64

ãĥĺ

0.63

acci

0.63

ractor

0.62

cel

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.