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

endas

-0.74

ascar

-0.72

hor

-0.71

ood

-0.71

Blocks

-0.69

arant

-0.68

among

-0.68

awed

-0.68

inside

-0.66

ab

-0.66

POSITIVE LOGITS

HAEL

0.65

ilee

0.64

 misses

0.63

 really

0.63

 discl

0.62

IENCE

0.62

BILITY

0.61

 McDonnell

0.60

 imag

0.60

ratulations

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.