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

FACE

-0.77

ogle

-0.71

retion

-0.66

glomer

-0.63

rolog

-0.62

DIV

-0.62

 ãĢĮ

-0.61

 Bild

-0.59

 devise

-0.59

 deceive

-0.57

POSITIVE LOGITS

throp

0.81

yards

0.75

loads

0.74

loaded

0.71

skinned

0.70

heard

0.69

arians

0.69

conv

0.68

appropri

0.68

yard

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.