INDEX

Explanations

the word "Hi"

oai_token-act-pair · gpt-3.5-turbo

occurrences of the greeting "Hi"

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

 Awakens

-0.85

å£«

-0.84

parts

-0.83

*/(

-0.81

 Hurricanes

-0.69

ãĥ¯

-0.68

 Rebell

-0.67

ioned

-0.66

arian

-0.66

 halves

-0.64

POSITIVE LOGITS

Hi

1.21

Hi

1.00

hi

0.86

hey

0.85

WARD

0.85

yip

0.82

bye

0.76

Fi

0.76

ixel

0.76

STER

0.73

Activations Density 0.005%

the word "Hi"

occurrences of the greeting "Hi"

No Comments

No Known Activations