INDEX

Explanations

mentions or references to LGBTQ+ individuals or topics

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "gay" in relation to rights and equality

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

âĵĺ

-0.81

KT

-0.80

Condition

-0.76

SEC

-0.72

Ct

-0.71

Reviewer

-0.71

ç«

-0.69

ICS

-0.69

nit

-0.68

ioxide

-0.67

POSITIVE LOGITS

 equality

1.05

gay

0.92

 bisexual

0.91

atri

0.89

glers

0.87

 gays

0.85

 Equality

0.85

sex

0.84

dar

0.82

 couples

0.82

Activations Density 0.017%

mentions or references to LGBTQ+ individuals or topics

mentions of the word "gay" in relation to rights and equality

No Comments

No Known Activations