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

ldom

-0.80

zanne

-0.78

folio

-0.76

cipled

-0.75

PDATE

-0.72

agine

-0.72

berus

-0.69

spir

-0.69

cffffcc

-0.69

nr

-0.68

POSITIVE LOGITS

éĸ

0.69

 Xavier

0.67

nance

0.64

Uz

0.63

 reprodu

0.63

ync

0.61

 explanatory

0.60

 Universal

0.58

éĢ

0.57

 Compton

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.