INDEX

Explanations

mentions of political figures and events

oai_token-act-pair · gpt-3.5-turbo

references to political figures and their connections

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

 Vaugh

-0.65

 experien

-0.62

ciating

-0.59

 horizont

-0.55

 disse

-0.55

 incorpor

-0.54

INAL

-0.54

 obser

-0.52

avascript

-0.51

 anecd

-0.50

POSITIVE LOGITS

 Nato

0.56

 Britain

0.55

 politic

0.53

 discredited

0.53

 threatened

0.51

 weakened

0.50

 billionaires

0.49

 subsid

0.49

zbollah

0.49

 Russia

0.48

Activations Density 1.626%

mentions of political figures and events

references to political figures and their connections

No Comments

No Known Activations