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

ETH

-0.68

 Elixir

-0.65

 Safety

-0.65

 Mental

-0.63

 outlines

-0.62

CE

-0.62

 Tome

-0.62

Song

-0.62

SI

-0.61

 Levels

-0.60

POSITIVE LOGITS

 prime

1.70

urn

1.43

prime

0.97

resa

0.89

inently

0.78

ndra

0.74

 cardinal

0.72

wered

0.72

odore

0.72

usalem

0.71

Activations Density 0.000%

No Known Activations

This feature has no known activations.