INDEX

Explanations

instances of the word "former" preceded by another word

oai_token-act-pair · gpt-3.5-turbo

references to entities that were previously significant or relevant

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

atown

-0.87

andise

-0.82

otle

-0.81

antics

-0.79

ourced

-0.78

achus

-0.76

anguage

-0.76

aris

-0.75

oked

-0.74

ahon

-0.73

POSITIVE LOGITS

 Yugoslavia

1.22

 Yugoslav

1.11

 Soviet

0.89

 occupant

0.78

 convict

0.76

 smoker

0.73

 USSR

0.72

 inmate

0.71

 soldier

0.70

 glory

0.70

Activations Density 0.031%

instances of the word "former" preceded by another word

references to entities that were previously significant or relevant

No Comments

No Known Activations