INDEX

Explanations

explicit mentions of laws or regulations

oai_token-act-pair · gpt-3.5-turbo

statements that begin with "states" or variations of it

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

iffe

-0.75

isoft

-0.65

ortun

-0.61

enic

-0.60

ãĤ¼

-0.59

 Neighbor

-0.59

 intrig

-0.58

 Flavoring

-0.57

 rains

-0.56

imet

-0.56

POSITIVE LOGITS

 unequivocally

1.05

 bluntly

0.94

 emphatically

0.86

 plainly

0.85

 goodbye

0.79

 "...

0.75

manship

0.74

rooms

0.71

quo

0.71

 categor

0.70

Activations Density 0.022%

explicit mentions of laws or regulations

statements that begin with "states" or variations of it

No Comments

No Known Activations