INDEX

Explanations

mentions of homophobic or anti-gay language, actions, or sentiments

oai_token-act-pair · gpt-3.5-turbo

references to anti-gay sentiments and legislation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

Dur

-0.70

 Chiefs

-0.70

 Grateful

-0.69

Bei

-0.67

 Planes

-0.65

 Solitaire

-0.64

externalActionCode

-0.63

 CentOS

-0.63

Nap

-0.62

ç«

-0.61

POSITIVE LOGITS

osexual

0.87

zilla

0.86

tons

0.83

nor

0.82

ega

0.81

erness

0.80

cation

0.79

gay

0.78

holes

0.75

ther

0.75

Activations Density 0.012%

mentions of homophobic or anti-gay language, actions, or sentiments

references to anti-gay sentiments and legislation

No Comments

No Known Activations