INDEX

Explanations

movie-related terms and references

oai_token-act-pair · gpt-3.5-turbo

mentions of movies

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

existent

-0.77

acia

-0.74

cffff

-0.72

adies

-0.71

odic

-0.71

omaly

-0.68

minded

-0.66

avis

-0.65

isol

-0.64

isms

-0.63

POSITIVE LOGITS

 theaters

1.22

 theater

1.17

 movie

1.16

 movies

1.08

movie

1.01

 theatre

1.01

Movie

0.99

goers

0.96

 Movie

0.95

 theat

0.95

Activations Density 0.022%

movie-related terms and references

mentions of movies

No Comments

No Known Activations