INDEX

Explanations

time-related terms such as days, hours, weeks, and months

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

llor

-0.80

ocard

-0.75

acted

-0.71

alach

-0.68

tarian

-0.67

atives

-0.64

VID

-0.62

 exception

-0.62

ICO

-0.61

apixel

-0.61

POSITIVE LOGITS

hift

0.93

entimes

0.85

peed

0.84

chool

0.82

ourcing

0.78

pring

0.78

 afterward

0.77

'-

0.77

 upon

0.77

pite

0.75

Activations Density 0.064%

time-related terms such as days, hours, weeks, and months

No Comments

No Known Activations