INDEX

Explanations

terms related to conspiracy theories and misinformation

oai_token-act-pair · gpt-3.5-turbo

terms associated with conspiracy theories and misinformation, particularly around climate change and fake news

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ktop

-0.78

eatures

-0.75

iosyn

-0.74

ugal

-0.73

illes

-0.70

mediately

-0.70

blance

-0.69

served

-0.65

 Duty

-0.65

 discharged

-0.64

POSITIVE LOGITS

 disinformation

1.14

 debunk

1.10

 debunked

1.10

 misinformation

1.03

icist

1.02

 propaganda

0.91

 hoax

0.89

 falsehood

0.87

 denial

0.87

ervative

0.87

Activations Density 0.332%

terms related to conspiracy theories and misinformation

terms associated with conspiracy theories and misinformation, particularly around climate change and fake news

No Comments

No Known Activations