INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

adow

-0.82

wi

-0.71

oda

-0.66

 parks

-0.63

jri

-0.63

 drought

-0.62

gha

-0.60

angers

-0.60

 Lama

-0.60

zza

-0.60

POSITIVE LOGITS

REF

0.98

MENTS

0.85

aughs

0.83

MENT

0.82

Critical

0.79

mort

0.78

DEF

0.76

Boo

0.74

Ell

0.72

Opt

0.71

Activations Density 0.000%

No Known Activations

This feature has no known activations.