INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Logged

-0.66

Bee

-0.65

Gas

-0.65

Phys

-0.64

 Barnett

-0.63

ACS

-0.63

 Welch

-0.62

 Luthor

-0.61

 Webb

-0.61

 Schne

-0.61

POSITIVE LOGITS

rams

0.78

emonic

0.76

ulum

0.74

undle

0.73

ilon

0.73

ossom

0.72

ough

0.72

anted

0.71

urdue

0.70

ilers

0.69

Activations Density 0.000%

No Known Activations

This feature has no known activations.