INDEX

Explanations

references to movie titles

oai_token-act-pair · gpt-3.5-turbo

titles of movies, TV shows, or video games

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

jri

-0.89

 tremend

-0.73

 skelet

-0.72

 bloss

-0.70

 denomin

-0.70

 advertis

-0.70

 opin

-0.68

 exhib

-0.67

yout

-0.67

yog

-0.67

POSITIVE LOGITS

 Shattered

1.04

 Rebirth

1.00

 Resurrection

1.00

 Trails

1.00

 Remastered

0.98

 Origins

0.98

 Siege

0.97

 Battle

0.96

 Deadly

0.96

 Conquest

0.96

Activations Density 0.038%

references to movie titles

titles of movies, TV shows, or video games

No Comments

No Known Activations