INDEX

Explanations

references to visual and media production terminology

oai_token-act-pair · gpt-3.5-turbo 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

Reloaded

-0.57

rown

-0.53

×ŀ

-0.52

)."

-0.52

ãĥ©ãĥ³

-0.51

 exha

-0.50

 doub

-0.50

 earnest

-0.50

ãĥ¡

-0.50

ãĥĳ

-0.50

POSITIVE LOGITS

↵Âł

0.76

 Copyright

0.76

?:

0.73

 illustrates

0.72

 Shutterstock

0.69

 Â©

0.67

 arrives

0.67

 demonstrates

0.66

 depicts

0.66

 appears

0.65

Activations Density 0.658%

references to visual and media production terminology

No Comments

No Known Activations