INDEX

Explanations

phrases related to presenting or emphasizing a specific point or belief

oai_token-act-pair · gpt-3.5-turbo

expressions that indicate assertion or belief

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 ÂŃ

-0.75

ÃĤ

-0.66

 withd

-0.57

 Composite

-0.56

 Vaugh

-0.55

-0.53

 âĶ

-0.52

âĢł

-0.51

 Azerb

-0.51

âĢ

-0.50

POSITIVE LOGITS

 fallacy

0.85

lessly

0.73

 doesnt

0.73

liness

0.70

 implies

0.67

wolves

0.65

 oneself

0.65

?),

0.64

 hath

0.64

ndra

0.63

Activations Density 0.934%

phrases related to presenting or emphasizing a specific point or belief

expressions that indicate assertion or belief

No Comments

No Known Activations