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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

[[

-0.70

Ghost

-0.66

Merit

-0.65

Prop

-0.62

NEC

-0.61

 heck

-0.59

GO

-0.59

Elect

-0.59

Tact

-0.58

vs

-0.57

POSITIVE LOGITS

nesday

0.89

his

0.89

 himself

0.89

umenthal

0.77

his

0.76

ousse

0.73

 hers

0.69

 eleph

0.69

 caut

0.68

imilar

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.