INDEX

Explanations

phrases related to public appearances or participation in events

oai_token-act-pair · gpt-3.5-turbo

references to the number of appearances made by individuals or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

apa

-0.79

ran

-0.74

stra

-0.70

jit

-0.68

efe

-0.67

tec

-0.67

BM

-0.66

 parcel

-0.64

rug

-0.64

adr

-0.63

POSITIVE LOGITS

 appearances

3.96

 appearance

2.54

earances

2.12

 Appearance

1.96

 outings

1.65

 sightings

1.39

 performances

1.34

 presentations

1.31

 cameo

1.30

 impressions

1.29

Activations Density 0.016%

phrases related to public appearances or participation in events

references to the number of appearances made by individuals or entities

No Comments

No Known Activations