INDEX

Explanations

phrases that prompt action or emphasize importance

oai_token-act-pair · gpt-3.5-turbo

references to objects or items being discussed

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

âĢ¢âĢ¢

-0.76

mire

-0.69

CCC

-0.65

âĺħâĺħ

-0.64

Iowa

-0.63

ILE

-0.62

chron

-0.62

Jo

-0.61

cgi

-0.60

execute

-0.59

POSITIVE LOGITS

atically

1.40

selves

1.39

 selves

1.34

atic

1.25

self

1.03

atics

0.84

MpServer

0.80

alian

0.79

abeth

0.77

 behav

0.77

Activations Density 0.127%

phrases that prompt action or emphasize importance

references to objects or items being discussed

No Comments

No Known Activations