INDEX

Explanations

mentions of TV shows and their episodes

oai_token-act-pair · gpt-3.5-turbo

references to television shows and their performances

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

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

yip

-1.01

Lenin

-0.87

imir

-0.87

ulic

-0.86

ventus

-0.84

 Hague

-0.83

fax

-0.81

veland

-0.80

unin

-0.75

userc

-0.73

POSITIVE LOGITS

episode

1.71

 episodes

1.70

 sitcom

1.65

 episode

1.60

Episode

1.55

 Season

1.54

 Emmy

1.52

 airing

1.46

 Episode

1.46

isodes

1.45

Activations Density 0.813%

mentions of TV shows and their episodes

references to television shows and their performances

No Comments

No Known Activations