INDEX

Explanations

words related to technical jargon and procedures in a formal context

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

ocl

-0.72

izen

-0.70

¬¼

-0.69

iple

-0.68

oll

-0.63

uber

-0.63

ire

-0.63

irie

-0.62

yon

-0.62

rio

-0.60

POSITIVE LOGITS

 namely

1.39

 although

1.34

 whereas

1.32

 albeit

1.30

 prompting

1.27

 thereby

1.27

 which

1.26

 including

1.25

 according

1.21

 implying

1.20

Activations Density 3.091%

words related to technical jargon and procedures in a formal context

No Comments

No Known Activations