INDEX

Explanations

words related to policy terms and expiration dates

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

ounding

-0.72

ilde

-0.70

akable

-0.68

anooga

-0.68

kell

-0.67

\":

-0.66

oly

-0.66

ortment

-0.66

opoulos

-0.66

asio

-0.63

POSITIVE LOGITS

 expire

1.27

 expired

1.20

 expiration

1.10

 expires

1.10

 prematurely

0.89

uate

0.84

 consecut

0.84

 leases

0.82

 timers

0.81

 curfew

0.80

Activations Density 0.028%

words related to policy terms and expiration dates

No Comments

No Known Activations