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

 pige

-0.87

 overfl

-0.76

pers

-0.75

efully

-0.65

cher

-0.65

dinand

-0.62

chery

-0.61

chers

-0.61

 alike

-0.61

 independ

-0.60

POSITIVE LOGITS

GREEN

0.75

wrote

0.66

lez

0.65

orah

0.64

Professor

0.64

 Fathers

0.63

cox

0.62

ãĥĥãĥĪ

0.62

cuts

0.62

Tes

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.