INDEX

Explanations

proper nouns related to politics, locations, and organizations

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

senal

-0.63

enegger

-0.57

foundland

-0.56

vertisement

-0.54

GGGGGGGG

-0.52

 looph

-0.52

retty

-0.51

 Leban

-0.50

 Bravo

-0.49

FontSize

-0.49

POSITIVE LOGITS

 cannot

0.70

 could

0.70

 deems

0.70

 would

0.66

 decides

0.66

 enters

0.66

 existed

0.65

 fails

0.64

had

0.64

 might

0.64

Activations Density 0.864%

proper nouns related to politics, locations, and organizations

No Comments

No Known Activations