INDEX

Explanations

mentions of podcasts and specific podcast episodes

oai_token-act-pair · gpt-3.5-turbo

references to wrestling podcasts and related discussions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

vernment

-0.83

 wrongly

-0.78

 labour

-0.75

bestos

-0.74

estate

-0.74

 Isis

-0.73

 kilometres

-0.70

 Nato

-0.70

lying

-0.69

 Identification

-0.68

POSITIVE LOGITS

Episode

1.13

 Episode

1.07

podcast

1.07

 Podcast

1.02

 podcast

0.97

 LIVE

0.96

 listener

0.94

PAX

0.94

 Patreon

0.93

isode

0.92

Activations Density 0.903%

mentions of podcasts and specific podcast episodes

references to wrestling podcasts and related discussions

No Comments

No Known Activations