INDEX

Explanations

names of people related to legal or ethical controversies

oai_token-act-pair · gpt-3.5-turbo

references to allegations or accusations involving individuals and misconduct

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

alg

-0.71

phabet

-0.70

 uniqueness

-0.70

eatures

-0.70

 clarity

-0.68

ainment

-0.66

ciation

-0.66

 reunion

-0.66

 partName

-0.65

ciating

-0.65

POSITIVE LOGITS

 improperly

1.50

 misled

1.43

 manipulated

1.39

 illegally

1.34

 knowingly

1.32

 inappropriately

1.32

 lied

1.30

 unlawfully

1.30

 fals

1.28

 violated

1.28

Activations Density 0.391%

names of people related to legal or ethical controversies

references to allegations or accusations involving individuals and misconduct

No Comments

No Known Activations