INDEX

Explanations

phrases related to numbers and statistics

oai_token-act-pair · gpt-3.5-turbo

legal terms and references related to crime and punishment

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ETA

-0.63

hack

-0.62

mx

-0.61

bill

-0.59

****************

-0.58

Ac

-0.57

Search

-0.57

pac

-0.57

Br

-0.56

Gall

-0.56

POSITIVE LOGITS

).[

0.86

rul

0.79

ļé

0.79

anwhile

0.74

)=(

0.72

 respectively

0.71

)).

0.70

 endif

0.70

SourceFile

0.69

ÃĥÃĤ

0.68

Activations Density 0.736%

phrases related to numbers and statistics

legal terms and references related to crime and punishment

No Comments

No Known Activations