INDEX

Explanations

words related to political positions or stances on various issues

oai_token-act-pair · gpt-3.5-turbo

references to political positions or stances

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ennes

-0.76

Champ

-0.72

Chel

-0.69

à¨

-0.67

agar

-0.64

INESS

-0.64

aniel

-0.63

@@@@

-0.62

flies

-0.62

anova

-0.61

POSITIVE LOGITS

 behalf

1.59

erous

0.99

 matters

0.90

yx

0.89

 occasion

0.87

 steroids

0.85

shore

0.83

 Capitol

0.81

eworld

0.81

eness

0.80

Activations Density 0.203%

words related to political positions or stances on various issues

references to political positions or stances

No Comments

No Known Activations