INDEX

Explanations

phrases related to causality and consequences

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

ovember

-0.66

tained

-0.65

senal

-0.58

idered

-0.58

affe

-0.58

ħĭ

-0.56

eatured

-0.56

arij

-0.56

ccording

-0.55

recent

-0.55

POSITIVE LOGITS

 theirs

1.01

 THEIR

0.80

 yours

0.79

 trillions

0.74

 shit

0.73

 THEY

0.71

 obey

0.70

 ours

0.70

 endlessly

0.69

 ALWAYS

0.68

Activations Density 1.105%

phrases related to causality and consequences

No Comments

No Known Activations