INDEX

Explanations

source code related to caching and data tokenization in programming

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

theless

-0.79

 gearing

-0.78

aeda

-0.72

 streng

-0.71

anthrop

-0.69

 acron

-0.69

 landsl

-0.69

 ancest

-0.68

 unavoid

-0.68

 jaws

-0.66

POSITIVE LOGITS

1.41

Value

1.30

()

1.29

();

1.25

Result

1.24

Header

1.23

&&

1.22

+=

1.21

Flags

1.21

Type

1.20

Activations Density 0.572%

source code related to caching and data tokenization in programming

No Comments

No Known Activations