INDEX

Explanations

keywords related to typography and formatting

oai_token-act-pair · gpt-3.5-turbo

the presence of a specific non-standard character

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Manit

-0.78

 Mous

-0.76

 Ferguson

-0.75

 USPS

-0.73

 Moss

-0.72

 shroud

-0.70

ACS

-0.69

 Tulsa

-0.68

Mul

-0.68

 Muss

-0.66

POSITIVE LOGITS

ľ

2.02

ł

1.72

Ķ

1.63

Ĳ

1.63

ª

1.57

Ŀ

1.52

ļ

1.49

¤

1.49

«

1.48

¬

1.48

Activations Density 0.095%

keywords related to typography and formatting

the presence of a specific non-standard character

No Comments

No Known Activations