INDEX

Explanations

mentions of leaked information or revelations

oai_token-act-pair · gpt-3.5-turbo

references to significant disclosures or revelations in political contexts

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

iculture

-0.87

otor

-0.86

 ç¥ŀ

-0.81

oba

-0.79

icultural

-0.79

Interstitial

-0.79

qus

-0.71

anguages

-0.71

icio

-0.71

oka

-0.69

POSITIVE LOGITS

 revelations

1.48

 leaked

1.46

 whistleblower

1.45

 disclosures

1.37

 allegations

1.29

 whistleblowers

1.28

 incrim

1.26

 disclosed

1.22

 leaks

1.21

 confidential

1.20

Activations Density 1.091%

mentions of leaked information or revelations

references to significant disclosures or revelations in political contexts

No Comments

No Known Activations