INDEX

Explanations

mentions of the term "Anti" followed by a single word

oai_token-act-pair · gpt-3.5-turbo

references to anti-related themes or concepts

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

hus

-0.78

 constrained

-0.65

 tantal

-0.64

 peek

-0.64

alus

-0.64

 calcul

-0.63

 deliberations

-0.62

 fortunate

-0.62

 snapping

-0.61

rows

-0.61

POSITIVE LOGITS

 Anti

3.57

Anti

2.64

 anti

1.56

anti

1.51

Ant

1.29

 Hate

1.23

 Counter

1.15

 antid

1.15

 Radical

1.13

 Corruption

1.09

Activations Density 0.013%

mentions of the term "Anti" followed by a single word

references to anti-related themes or concepts

No Comments

No Known Activations