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

phrine

-0.75

 Restaur

-0.71

 Palmer

-0.69

 Neurolog

-0.68

 Protein

-0.64

 Goose

-0.63

 Ethiop

-0.62

 Stephenson

-0.61

 Gast

-0.61

 FreeBSD

-0.61

POSITIVE LOGITS

ebin

0.87

 myster

0.77

areth

0.74

 srfAttach

0.71

 explan

0.71

antis

0.70

endi

0.68

oola

0.67

eared

0.66

nikov

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.