INDEX

Explanations

mentions of positions or rankings

oai_token-act-pair · gpt-3.5-turbo

colons and numerical data or statistics related to rankings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

bered

-0.76

 tremend

-0.75

omething

-0.74

inement

-0.72

vre

-0.70

uggest

-0.70

uve

-0.69

izons

-0.69

schild

-0.67

alist

-0.65

POSITIVE LOGITS

 Provided

0.96

???

0.93

 Unknown

0.92

TBD

0.91

Bye

0.85

TBA

0.83

 None

0.79

 Same

0.78

 Miscellaneous

0.76

 Nope

0.75

Activations Density 0.085%

mentions of positions or rankings

colons and numerical data or statistics related to rankings

No Comments

No Known Activations