INDEX

Explanations

phrases related to unidentified or mysterious entities

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "unknown."

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

ousse

-0.66

Ħ¢

-0.66

hao

-0.65

ARE

-0.65

PRESS

-0.64

psc

-0.63

jad

-0.62

ards

-0.61

eret

-0.61

Tweet

-0.60

POSITIVE LOGITS

 unknown

3.93

unknown

2.50

 Unknown

2.39

 undisclosed

1.96

 unspecified

1.96

 unidentified

1.89

Unknown

1.86

 unseen

1.81

 uncertain

1.73

 unclear

1.63

Activations Density 0.025%

phrases related to unidentified or mysterious entities

references to the concept of "unknown."

No Comments

No Known Activations