INDEX

Explanations

words related to genuine or unaltered things

oai_token-act-pair · gpt-3.5-turbo

references to authenticity and its variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

row

-0.89

wind

-0.77

 Waste

-0.74

rows

-0.74

rough

-0.71

gall

-0.70

Column

-0.69

Wind

-0.69

kel

-0.68

urden

-0.65

POSITIVE LOGITS

 authentic

3.81

 Authent

2.48

authent

2.39

 authenticity

2.38

 authent

2.26

Authent

1.96

 genuine

1.76

 authenticated

1.62

 AUTH

1.39

 legit

1.38

Activations Density 0.015%

words related to genuine or unaltered things

references to authenticity and its variations

No Comments

No Known Activations