INDEX

Explanations

professionals or experts in various fields

oai_token-act-pair · gpt-3.5-turbo

noun phrases related to rankings or status within various fields

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

rontal

-0.73

tnc

-0.72

ãĥĥ

-0.72

 srfAttach

-0.71

çīĪ

-0.70

rocal

-0.69

 guiActiveUn

-0.69

shadow

-0.65

uish

-0.63

ritz

-0.62

POSITIVE LOGITS

 EVER

1.38

 ever

1.35

 imaginable

1.31

ever

1.00

 anywhere

0.98

Ever

0.94

 alive

0.92

 Ever

0.90

 worldwide

0.84

 globally

0.81

Activations Density 0.242%

professionals or experts in various fields

noun phrases related to rankings or status within various fields

No Comments

No Known Activations