INDEX

Explanations

references to individuals named "Edward"

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Edward."

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

 carbs

-0.73

minimum

-0.68

plan

-0.68

itiz

-0.66

Ig

-0.65

carb

-0.63

sor

-0.63

stick

-0.62

 reservations

-0.62

 plasma

-0.62

POSITIVE LOGITS

 Edward

3.53

Edward

2.89

 Edwin

1.57

 Snowden

1.48

 Henry

1.46

 William

1.44

 Frederick

1.43

 Edgar

1.34

 Elizabeth

1.34

 Geoffrey

1.33

Activations Density 0.012%

references to individuals named "Edward"

mentions of the name "Edward."

No Comments

No Known Activations