INDEX

Explanations

mentions of syndicated radio shows or programs

oai_token-act-pair · gpt-3.5-turbo

terms related to syndication and syndicates

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ghazi

-0.75

FUL

-0.71

 Downing

-0.68

DOE

-0.67

Lot

-0.66

boarding

-0.66

BALL

-0.65

tesy

-0.65

dash

-0.64

ãģĤ

-0.64

POSITIVE LOGITS

 synd

1.46

icate

1.31

icates

1.18

icated

1.17

ication

1.09

 mosqu

1.02

 Synd

0.98

icator

0.98

emic

0.93

icating

0.91

Activations Density 0.004%

mentions of syndicated radio shows or programs

terms related to syndication and syndicates

No Comments

No Known Activations