INDEX

Explanations

personal pronouns followed by verbs indicating personal actions

oai_token-act-pair · gpt-3.5-turbo

the placeholder token indicating the end of a document

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Georgian

-0.57

 selves

-0.53

 srfAttach

-0.52

ãĥĥãĥī

-0.49

 bodies

-0.49

 branches

-0.48

farious

-0.48

 prompting

-0.47

ifiable

-0.47

 Plaint

-0.47

POSITIVE LOGITS

'm

1.14

stanbul

1.03

've

0.95

EEE

0.94

nex

0.91

am

0.86

 suppose

0.86

ronic

0.86

 adore

0.86

zzy

0.85

Activations Density 0.197%

personal pronouns followed by verbs indicating personal actions

the placeholder token indicating the end of a document

No Comments

No Known Activations