INDEX

Explanations

images with water-related content

oai_token-act-pair · gpt-3.5-turbo

separators or slashes used in image or photo citations

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

 debates

-0.84

 demol

-0.84

 witnessing

-0.82

 sweeping

-0.81

 favour

-0.80

 pronoun

-0.79

 amounts

-0.78

 clam

-0.77

 infiltr

-0.76

 remed

-0.75

POSITIVE LOGITS

Flickr

1.94

Getty

1.78

Wik

1.77

YouTube

1.74

AFP

1.68

Courtesy

1.57

Twitter

1.56

Reuters

1.53

Tumblr

1.52

EPA

1.51

Activations Density 0.047%

images with water-related content

separators or slashes used in image or photo citations

No Comments

No Known Activations