INDEX

Explanations

mentions of individuals or groups of people

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "someone" and its variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

tnc

-0.77

DOS

-0.73

effect

-0.72

osterone

-0.71

ory

-0.71

Rom

-0.70

UV

-0.68

ean

-0.67

AIN

-0.66

CDC

-0.66

POSITIVE LOGITS

 else

1.81

Else

1.27

else

1.17

 Else

1.10

 forgot

0.97

 stole

0.96

 clicked

0.91

 screwed

0.86

 yelled

0.83

 threw

0.82

Activations Density 0.041%

mentions of individuals or groups of people

occurrences of the word "someone" and its variations

No Comments

No Known Activations