INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

proper names or significant titles within the text

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

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

 âĢº

-0.90

 Ying

-0.73

 Hubble

-0.71

 waterfall

-0.68

 adop

-0.67

 Roku

-0.67

 ãĢĮ

-0.66

chart

-0.65

âĢ

-0.65

Recomm

-0.64

POSITIVE LOGITS

ulz

2.61

ooter

1.78

 Emma

1.54

ooters

1.44

 masked

1.34

ooting

1.31

nikov

1.15

olver

1.01

kinson

1.01

lag

0.98

Activations Density 0.027%

proper nouns

proper names or significant titles within the text

No Comments

No Known Activations