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

orum

-0.73

 damp

-0.66

 cursed

-0.63

odge

-0.63

alter

-0.61

ulz

-0.61

 wool

-0.61

omer

-0.60

etts

-0.60

athe

-0.60

POSITIVE LOGITS

BLIC

0.72

assador

0.68

GER

0.65

enary

0.63

VICE

0.63

Tank

0.62

agascar

0.62

ortium

0.62

FLAG

0.61

ACP

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.