INDEX

Explanations

personal pronouns and gendered nouns

oai_token-act-pair · gpt-3.5-turbo

references to male and female characters

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

grave

-0.76

assing

-0.75

stellar

-0.66

ylon

-0.65

igmatic

-0.64

kefeller

-0.63

 Observatory

-0.62

Ĥª

-0.61

irm

-0.61

cgi

-0.60

POSITIVE LOGITS

mos

0.95

 Majesty

0.94

'll

0.93

 didn

0.86

'd

0.85

 knew

0.84

 wanted

0.82

 knows

0.81

 didnt

0.80

 hates

0.80

Activations Density 0.237%

personal pronouns and gendered nouns

references to male and female characters

No Comments

No Known Activations