INDEX

Explanations

phrases related to action and activity

oai_token-act-pair · gpt-3.5-turbo

punctuation and its frequency in different contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.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

uclear

-0.97

vernment

-0.97

Netflix

-0.94

uterte

-0.92

tarian

-0.92

FactoryReloaded

-0.90

Nintendo

-0.90

TRUMP

-0.90

obyl

-0.89

Saudi

-0.87

POSITIVE LOGITS

 teamwork

0.82

 Specialist

0.73

 each

0.71

 Founder

0.71

 these

0.70

 Each

0.68

 beginners

0.68

 frontline

0.67

 skilled

0.67

 Winning

0.67

Activations Density 0.710%

phrases related to action and activity

punctuation and its frequency in different contexts

No Comments

No Known Activations