INDEX

Explanations

the repeated mention of the term "An" or its variations in different contexts

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

occurrences of the word "An" followed by another word

oai_token-act-pair · gpt-3.5-turbo

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

âĸº

-0.64

 Continued

-0.62

 jeopardy

-0.61

 minus

-0.60

 fastball

-0.59

 sides

-0.59

 backwards

-0.57

 wink

-0.57

yes

-0.56

 Solitaire

-0.55

POSITIVE LOGITS

omaly

1.39

alogue

1.37

alys

1.31

ubis

1.29

nette

1.23

omal

1.23

alyses

1.18

notation

1.18

terior

1.16

esthesia

1.14

Activations Density 0.022%

the repeated mention of the term "An" or its variations in different contexts

occurrences of the word "An" followed by another word

No Comments

No Known Activations