INDEX

Explanations

phrases referring to someone requesting to remain anonymous or unidentified

oai_token-act-pair · gpt-3.5-turbo

references to anonymity or confidentiality in statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

Ļ½

-0.69

 pedest

-0.66

chrome

-0.61

 wagon

-0.61

 bandwagon

-0.59

 Tend

-0.59

ãĢĲ

-0.58

 sights

-0.58

 Goddess

-0.55

 playbook

-0.54

POSITIVE LOGITS

 anonymity

1.34

 pseudonym

0.95

 privacy

0.82

 confidentiality

0.82

 surn

0.80

 anonymous

0.77

 pseud

0.75

 anonymously

0.75

 surname

0.73

 attribution

0.73

Activations Density 0.204%

phrases referring to someone requesting to remain anonymous or unidentified

references to anonymity or confidentiality in statements

No Comments

No Known Activations