INDEX

Explanations

names of TV shows and movies

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names related to characters, shows, or notable entities in entertainment

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 scattering

-0.76

 Negro

-0.69

 unwanted

-0.69

 powerless

-0.68

 salv

-0.66

 helpless

-0.66

 unexpl

-0.66

 unintended

-0.65

 mushroom

-0.65

 strugg

-0.65

POSITIVE LOGITS

ï¸ı

1.27

Â¯

1.13

ï¸

1.13

âĶģ

0.92

å§«

0.89

Tea

0.85

âĻ

0.82

tm

0.81

ÃĽ

0.78

âĢ

0.78

Activations Density 0.160%

names of TV shows and movies

proper nouns, particularly names related to characters, shows, or notable entities in entertainment

No Comments

No Known Activations