INDEX

Explanations

words associated with contrast or additional information

oai_token-act-pair · gpt-3.5-turbo

the word "also" and its variations indicating additional information or factors

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ASE

-0.72

UGE

-0.69

enter

-0.69

agus

-0.68

ILLE

-0.66

 Lash

-0.64

owship

-0.64

 Nine

-0.64

 Ange

-0.63

IT

-0.62

POSITIVE LOGITS

 includ

0.88

 includes

0.81

roph

0.79

 occasionally

0.78

 optionally

0.76

 incorporates

0.74

 happens

0.73

 contribute

0.72

 anecd

0.72

 contributes

0.71

Activations Density 0.112%

words associated with contrast or additional information

the word "also" and its variations indicating additional information or factors

No Comments

No Known Activations