INDEX

Explanations

phrases related to the act of reading or consuming information

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 signifies

-0.76

 Reduce

-0.74

 Ensure

-0.73

udeau

-0.73

 Previous

-0.73

pedia

-0.70

 Includes

-0.70

xton

-0.69

 Implementation

-0.68

 denotes

-0.68

POSITIVE LOGITS

 willing

1.22

 afraid

1.17

 interested

1.14

 gonna

1.10

 clam

1.09

 able

1.07

 aware

1.06

 hungry

1.03

 unaware

1.02

 eager

1.01

Activations Density 0.178%

phrases related to the act of reading or consuming information

No Comments

No Known Activations