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

 Vampire

-0.74

 Decision

-0.74

 Happiness

-0.71

oneliness

-0.66

ossip

-0.66

appiness

-0.66

acle

-0.65

isons

-0.63

alogue

-0.62

onement

-0.62

POSITIVE LOGITS

ItemTracker

0.90

 tremend

0.88

dos

0.79

everal

0.74

 confir

0.74

oute

0.71

Ĥİ

0.71

 sled

0.69

isphere

0.68

rontal

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.