INDEX

Explanations

phrases related to bias and skewed data

oai_token-act-pair · gpt-3.5-turbo

terms related to bias or distortion in data representation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

duct

-0.73

aido

-0.70

clamation

-0.69

ain

-0.69

pressed

-0.66

perial

-0.66

foundation

-0.65

Esc

-0.65

mediate

-0.65

ension

-0.64

POSITIVE LOGITS

ske

1.30

 skew

1.05

balls

0.73

 Dise

0.73

 pies

0.72

 variance

0.70

 skewed

0.69

ering

0.67

 Returning

0.67

ws

0.65

Activations Density 0.012%

phrases related to bias and skewed data

terms related to bias or distortion in data representation

No Comments

No Known Activations