INDEX

Explanations

phrases related to actions taking place at specific points in time

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "during."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

AAAA

-0.63

AAA

-0.57

 Godd

-0.55

cott

-0.55

rane

-0.55

utable

-0.55

 reckon

-0.54

rils

-0.54

 Tale

-0.52

 abstract

-0.52

POSITIVE LOGITS

 during

2.85

during

2.46

 During

1.84

During

1.79

 throughout

1.47

 afterward

1.33

 whilst

1.33

 shortly

1.29

 while

1.24

 afterwards

1.22

Activations Density 0.047%

phrases related to actions taking place at specific points in time

occurrences of the word "during."

No Comments

No Known Activations