INDEX

Explanations

specific numerical patterns

oai_token-act-pair · gpt-3.5-turbo

numerical data and references to figures in the text

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

DragonMagazine

-0.76

-0.66

paralle

-0.64

 Marathon

-0.60

 defamation

-0.59

EngineDebug

-0.59

 disqualified

-0.57

userc

-0.56

 blinding

-0.56

 disqual

-0.56

POSITIVE LOGITS

1.07

xi

1.03

1.00

qi

0.99

XL

0.99

0.98

aq

0.95

0.94

0.93

Activations Density 0.193%

specific numerical patterns

numerical data and references to figures in the text

No Comments

No Known Activations