INDEX

Explanations

terms related to accuracy and inauguration

oai_token-act-pair · gpt-3.5-turbo

terms related to accuracy and related concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 lift

-0.67

 whisk

-0.65

 solicitation

-0.63

 lines

-0.63

 Sect

-0.62

boo

-0.60

bra

-0.59

craft

-0.59

cor

-0.59

 wings

-0.59

POSITIVE LOGITS

urate

4.62

urated

2.52

uracy

2.32

uration

2.22

urations

1.68

urable

1.41

ur

1.26

ostic

1.20

uristic

1.20

ural

1.20

Activations Density 0.019%

terms related to accuracy and inauguration

terms related to accuracy and related concepts

No Comments

No Known Activations