INDEX

Explanations

exciting and positive expressions

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

 subsistence

-0.70

govtrack

-0.69

ourses

-0.69

obook

-0.68

 predomin

-0.68

icipated

-0.66

 precip

-0.65

 gradually

-0.64

ourse

-0.64

 principally

-0.63

POSITIVE LOGITS

!!!!!

1.38

!!!!

1.33

!!!

1.31

!!

1.29

!!!!!!!!

1.23

oooo

1.19

oooooooo

1.15

ðŁĺ

1.09

?!

1.09

 haha

1.08

Activations Density 2.442%

exciting and positive expressions

No Comments

No Known Activations