INDEX

Explanations

email-related phrases, specifically mentions about updates and offers

oai_token-act-pair · gpt-3.5-turbo

references to updates

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

 Galile

-0.68

Rim

-0.58

eur

-0.56

 idol

-0.54

Arg

-0.52

pedia

-0.52

 Goose

-0.52

ries

-0.51

 ammonia

-0.51

agate

-0.51

POSITIVE LOGITS

ilver

0.70

weet

0.68

hooting

0.66

iberal

0.65

ettings

0.65

âĺħ

0.64

ooters

0.64

ilty

0.63

daily

0.63

scl

0.63

Activations Density 0.021%

email-related phrases, specifically mentions about updates and offers

references to updates

No Comments

No Known Activations