INDEX

Explanations

mentions of women addressed with a title such as "Mrs" with a low activation for the "Mrs." variant

oai_token-act-pair · gpt-3.5-turbo

occurrences of the title "Mrs" before names

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

umbn

-0.89

anwhile

-0.80

 Archdemon

-0.78

undo

-0.77

ioch

-0.75

osta

-0.73

amaru

-0.72

adiator

-0.71

ahime

-0.70

 skelet

-0.68

POSITIVE LOGITS

 Clinton

0.84

 Claus

0.82

Mae

0.77

 Hillary

0.77

 Thatcher

0.75

 Theresa

0.74

 Jane

0.73

rha

0.71

 Diane

0.71

 Bride

0.70

Activations Density 0.014%

mentions of women addressed with a title such as "Mrs" with a low activation for the "Mrs." variant

occurrences of the title "Mrs" before names

No Comments

No Known Activations