INDEX

Explanations

terms related to mythology, particularly names ending with "og."

oai_token-act-pair · gpt-3.5-turbo

mentions of specific fictional characters or entities, particularly those associated with a narrative or prophecy context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 delegate

-0.75

 Sinai

-0.75

 delegates

-0.67

 Benjamin

-0.63

 transmitter

-0.62

 sterling

-0.61

 Rece

-0.60

Øª

-0.59

 Daniels

-0.59

 Blackburn

-0.59

POSITIVE LOGITS

og

4.33

ogs

2.56

ogi

1.84

ogging

1.83

OG

1.79

ogy

1.75

ogged

1.73

ogie

1.71

ogg

1.66

oga

1.63

Activations Density 0.019%

terms related to mythology, particularly names ending with "og."

mentions of specific fictional characters or entities, particularly those associated with a narrative or prophecy context

No Comments

No Known Activations