INDEX

Explanations

mentions of the media

oai_token-act-pair · gpt-3.5-turbo

mentions of the media

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

vasive

-0.82

thens

-0.73

 Parenthood

-0.71

ajor

-0.70

cker

-0.69

Sap

-0.68

Tec

-0.65

erville

-0.63

 Wrath

-0.63

adian

-0.63

POSITIVE LOGITS

media

1.06

 outlets

1.02

 media

1.00

eval

1.00

Media

0.93

 outlet

0.86

wiki

0.83

mog

0.83

medi

0.74

 plurality

0.73

Activations Density 0.031%

mentions of the media

mentions of the media

No Comments

No Known Activations