INDEX

Explanations

phrases mentioning documents or reports, especially those related to scandals or leaks

oai_token-act-pair · gpt-3.5-turbo

references to the Panama Papers and associated leaks

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

nec

-0.74

 Lansing

-0.70

UGH

-0.70

VIS

-0.68

 Baal

-0.67

CONCLUS

-0.65

aud

-0.65

GY

-0.64

east

-0.62

^^^^

-0.62

POSITIVE LOGITS

 Papers

1.51

 Paper

0.93

Paper

0.80

Leaks

0.77

0.76

 Archive

0.74

 Documents

0.73

breakers

0.72

mith

0.70

 Pengu

0.69

Activations Density 0.014%

phrases mentioning documents or reports, especially those related to scandals or leaks

references to the Panama Papers and associated leaks

No Comments

No Known Activations