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

*****

-0.84

****

-0.77

externalActionCode

-0.75

 à¨

-0.74

 Sabha

-0.71

 Afgh

-0.70

rawdownloadcloneembedreportprint

-0.69

Ü

-0.68

..................

-0.67

POSITIVE LOGITS

ansson

0.72

otom

0.70

 bund

0.69

rils

0.69

facts

0.66

reatment

0.66

olia

0.65

haar

0.63

obal

0.63

olkien

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.