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

pire

-0.74

riction

-0.68

pires

-0.68

 Django

-0.64

 Summoner

-0.62

 Pepe

-0.60

 friction

-0.59

venge

-0.58

 wrinkles

-0.57

rgb

-0.57

POSITIVE LOGITS

mate

1.10

mates

1.09

 mates

0.92

 mate

0.89

 Pats

0.79

iband

0.75

 Canad

0.69

ochond

0.69

 Burgess

0.69

upon

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.