INDEX

Explanations

references to different sections or categories within a text

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

 ende

-0.70

 Calder

-0.68

 amnesty

-0.67

 airst

-0.66

 unsc

-0.66

 tram

-0.66

 clin

-0.65

 mant

-0.64

 Norm

-0.64

hement

-0.62

POSITIVE LOGITS

https

1.15

http

1.15

taboola

1.01

Interstitial

1.00

BuyableInstoreAndOnline

0.99

SCP

0.98

Python

0.95

Community

0.92

Pinterest

0.91

www

0.89

Activations Density 0.084%

references to different sections or categories within a text

No Comments

No Known Activations