INDEX

Explanations

mentions of political affiliations, specifically liberals

oai_token-act-pair · gpt-3.5-turbo

references to liberals in a political context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

lying

-0.71

teen

-0.69

Temperature

-0.67

Sul

-0.66

aring

-0.65

Ther

-0.65

yss

-0.63

condition

-0.62

 Occupations

-0.62

ving

-0.61

POSITIVE LOGITS

ervatives

1.07

aurus

1.00

ervative

0.97

enos

0.78

 Liberals

0.77

umenthal

0.76

 liberals

0.76

arians

0.73

hift

0.72

paces

0.72

Activations Density 0.011%

mentions of political affiliations, specifically liberals

references to liberals in a political context

No Comments

No Known Activations