INDEX

Explanations

political references, specifically related to political parties

oai_token-act-pair · gpt-3.5-turbo

references to political parties and their affiliations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

antha

-0.74

inventoryQuantity

-0.72

Witness

-0.69

perties

-0.67

 journalistic

-0.64

judicial

-0.64

Sense

-0.63

Ļ

-0.63

profile

-0.62

¥ŀ

-0.61

POSITIVE LOGITS

ervative

0.97

ervatives

0.85

 wisely

0.83

 themselves

0.81

linger

0.79

0.78

layer

0.78

hift

0.74

aurus

0.74

heet

0.73

Activations Density 0.100%

political references, specifically related to political parties

references to political parties and their affiliations

No Comments

No Known Activations