INDEX

Explanations

mentions of leaked or hacked documents and sensitive personal information

oai_token-act-pair · gpt-3.5-turbo

references to data leaks and sensitive information

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

 Rally

-0.79

 masc

-0.77

spons

-0.71

vantage

-0.71

 Chaser

-0.70

Coach

-0.69

 relegation

-0.69

 Racer

-0.69

 rode

-0.69

undai

-0.68

POSITIVE LOGITS

 trove

1.47

 transcripts

1.43

 documents

1.43

 records

1.41

 archives

1.37

 files

1.36

 archive

1.26

 databases

1.22

 data

1.20

 recordings

1.19

Activations Density 0.788%

mentions of leaked or hacked documents and sensitive personal information

references to data leaks and sensitive information

No Comments

No Known Activations