INDEX

Explanations

phrases related to specific names or locations

oai_token-act-pair · gpt-3.5-turbo

references to specific organizations, people, or concepts related to LGBTQ+ history and cultural topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Ferr

-0.49

 Schw

-0.46

 warrants

-0.44

href

-0.44

orem

-0.44

rett

-0.44

 Conrad

-0.43

 Prometheus

-0.42

KS

-0.42

 disg

-0.42

POSITIVE LOGITS

heed

0.63

vironment

0.59

utenant

0.58

 Klux

0.56

abama

0.54

racuse

0.54

\":

0.54

 Flavoring

0.53

chwitz

0.52

uthor

0.52

Activations Density 0.369%

phrases related to specific names or locations

references to specific organizations, people, or concepts related to LGBTQ+ history and cultural topics

No Comments

No Known Activations