INDEX

Explanations

proper nouns ending in 'se'

oai_token-act-pair · gpt-3.5-turbo

the word "Asian" and its various forms or contexts

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

è¦ļéĨĴ

-0.76

owl

-0.69

 hawk

-0.69

 hoops

-0.67

 shorth

-0.65

 scheduling

-0.63

 wildfire

-0.63

iants

-0.61

 liking

-0.61

 fences

-0.61

POSITIVE LOGITS

eker

1.27

gger

1.01

venth

1.00

gments

0.97

vere

0.97

ldom

0.92

vier

0.91

ve

0.90

vent

0.89

xy

0.89

Activations Density 0.021%

proper nouns ending in 'se'

the word "Asian" and its various forms or contexts

No Known Activations