INDEX

Explanations

phrases that indicate the quality or ranking of something

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

arist

-0.82

athered

-0.81

umat

-0.80

ĸļ

-0.77

redo

-0.75

areth

-0.75

ipient

-0.74

reon

-0.73

cia

-0.71

acerb

-0.71

POSITIVE LOGITS

 aspects

0.75

 places

0.75

 moments

0.75

 niche

0.72

 ways

0.71

 smartphone

0.71

 parts

0.71

 edible

0.70

 possible

0.70

Favorite

0.69

Activations Density 0.075%

phrases that indicate the quality or ranking of something

No Comments

No Known Activations