INDEX

Explanations

information related to news articles, political figures, and media reporting

oai_token-act-pair · gpt-3.5-turbo

discussions related to political news and debates

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

hess

-0.68

nexus

-0.66

////////////////////////////////

-0.61

 realise

-0.60

imei

-0.60

cube

-0.59

 fitt

-0.59

oire

-0.58

cyclop

-0.58

POSITIVE LOGITS

 Obamacare

1.39

 Donald

1.38

 Benghazi

1.36

 Hillary

1.33

 ObamaCare

1.29

 Trump

1.28

 Kavanaugh

1.25

 Charlottesville

1.25

 Ivanka

1.23

Hillary

1.22

Activations Density 0.790%

information related to news articles, political figures, and media reporting

discussions related to political news and debates

No Comments

No Known Activations