INDEX

Explanations

instances of a sequential occurrence of events or items

oai_token-act-pair · gpt-3.5-turbo

phrases that reference a series or sequence of events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

uca

-0.87

ebus

-0.76

ptroller

-0.75

eri

-0.73

Sport

-0.72

someone

-0.72

UME

-0.71

enery

-0.71

WAY

-0.71

etus

-0.71

POSITIVE LOGITS

 events

1.16

 setbacks

1.01

 posts

0.98

 meetings

0.96

 tweets

0.96

 announcements

0.94

 crises

0.92

 compromises

0.92

 transformations

0.91

 episodes

0.91

Activations Density 0.062%

instances of a sequential occurrence of events or items

phrases that reference a series or sequence of events

No Comments

No Known Activations