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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

anding

-0.80

 Rowling

-0.77

icides

-0.73

anu

-0.72

que

-0.66

ults

-0.66

ooks

-0.66

ovic

-0.65

illard

-0.64

iques

-0.64

POSITIVE LOGITS

Syn

0.85

Ay

0.81

terday

0.78

\\\\

0.76

Textures

0.75

trial

0.73

ãĥ©ãĥ³

0.72

\\\\\\\\

0.70

ORGE

0.69

íķ

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.