INDEX

Explanations

mentions of political figures or topics

oai_token-act-pair · gpt-3.5-turbo

references to political themes or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

LEASE

-0.74

ENDED

-0.70

VIEW

-0.70

lighting

-0.69

noon

-0.67

 Carbuncle

-0.66

upon

-0.65

ounty

-0.65

RIS

-0.65

 pity

-0.65

POSITIVE LOGITS

 Polit

1.30

ifact

1.08

icians

1.06

ician

0.94

icial

0.88

ically

0.87

 paran

0.85

 correctness

0.83

Polit

0.80

 tabl

0.78

Activations Density 0.006%

mentions of political figures or topics

references to political themes or entities

No Comments

No Known Activations