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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

hel

-0.66

pit

-0.65

 srfAttach

-0.65

 Rabbit

-0.64

 genetics

-0.63

 Marilyn

-0.63

 torso

-0.62

 Mast

-0.62

 Frankenstein

-0.62

 Sabbath

-0.61

POSITIVE LOGITS

avorite

0.77

ãĥ©ãĥ³

0.76

sbm

0.75

apter

0.74

ardless

0.73

owell

0.72

agements

0.69

ontent

0.69

NB

0.69

Ú

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.