INDEX

Explanations

phrases or words related to legal or formal citations

oai_token-act-pair · gpt-3.5-turbo

references to citations within the text

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

yss

-0.76

akia

-0.76

nut

-0.73

yles

-0.72

enthal

-0.71

aido

-0.71

ntil

-0.71

olen

-0.71

nu

-0.70

 independ

-0.69

POSITIVE LOGITS

 citation

1.75

 citations

1.53

 Citation

1.10

 cite

0.92

 cited

0.91

 footnote

0.89

 Clicker

0.88

ibli

0.80

 attribution

0.77

DragonMagazine

0.75

Activations Density 0.009%

phrases or words related to legal or formal citations

references to citations within the text

No Comments

No Known Activations