INDEX

Explanations

mentions of LGBTQ-related terms, specifically "gay"

oai_token-act-pair · gpt-3.5-turbo

mentions of gay identity and associated rights

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

è¦ļéĨĴ

-0.75

arily

-0.71

urers

-0.70

Dispatch

-0.70

Condition

-0.67

ULE

-0.67

 sidx

-0.67

âĵĺ

-0.67

acqu

-0.67

 Manufacturer

-0.66

POSITIVE LOGITS

atri

0.99

 couples

0.94

dar

0.93

 marriage

0.92

 porn

0.87

lord

0.85

 slurs

0.85

ety

0.84

 bashing

0.83

ening

0.82

Activations Density 0.034%

mentions of LGBTQ-related terms, specifically "gay"

mentions of gay identity and associated rights

No Comments

No Known Activations