INDEX

Explanations

podcast-related words and phrases

oai_token-act-pair · gpt-3.5-turbo

references to podcasts

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

axy

-0.75

ju

-0.73

arde

-0.72

illard

-0.70

metics

-0.68

lda

-0.66

pes

-0.65

ple

-0.64

enes

-0.63

bang

-0.63

POSITIVE LOGITS

 listeners

1.13

 podcasts

1.01

 listener

0.98

 podcast

0.93

 episodes

0.89

podcast

0.89

 episode

0.88

odcast

0.87

 subscriptions

0.85

Episode

0.85

Activations Density 0.035%

podcast-related words and phrases

references to podcasts

No Comments

No Known Activations