INDEX

Explanations

uppercase letters or acronyms in the document

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

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

 ILCS

-0.72

-+-+

-0.69

 stru

-0.66

ãĤ©

-0.66

 unsupported

-0.65

 Sweeney

-0.65

 Bere

-0.64

à¤

-0.63

\/\/

-0.63

 TABLE

-0.63

POSITIVE LOGITS

tg

0.83

yu

0.83

tarian

0.79

Fi

0.79

idian

0.76

vP

0.75

ymes

0.74

dL

0.74

yne

0.72

Ha

0.72

Activations Density 0.064%

uppercase letters or acronyms in the document

No Comments

No Known Activations