INDEX

Explanations

phrases related to time spent or invested in various activities or situations

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

named

-0.72

eller

-0.63

deal

-0.60

 Parade

-0.59

 Cosponsors

-0.59

idence

-0.58

 assassinated

-0.58

ĪĴ

-0.57

tain

-0.57

 Cricket

-0.56

POSITIVE LOGITS

itures

1.01

 eternity

0.97

lav

0.90

iture

0.87

IPM

0.83

 untold

0.83

 trillions

0.81

 time

0.80

 precious

0.80

 hours

0.79

Activations Density 4.388%

phrases related to time spent or invested in various activities or situations

No Comments

No Known Activations