INDEX

Explanations

percentages and numbers related to statistics and data

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

lay

-0.65

folk

-0.63

ramid

-0.63

atem

-0.60

amaru

-0.58

ilic

-0.58

vati

-0.58

imaru

-0.56

elf

-0.56

Fram

-0.55

POSITIVE LOGITS

1.47

%-

1.36

 percent

1.35

%"

1.35

1.26

%,

1.25

%;

1.24

%.

1.23

%:

1.22

th

1.16

Activations Density 1.600%

percentages and numbers related to statistics and data

No Comments

No Known Activations