INDEX

Explanations

news-related terms like "published" and "released."

oai_token-act-pair · gpt-3.5-turbo

instances of published or released information, typically from authoritative sources

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

abiding

-0.76

haven

-0.68

soc

-0.66

asse

-0.66

externalActionCode

-0.65

-|

-0.61

intend

-0.61

soever

-0.60

stall

-0.59

ooters

-0.59

POSITIVE LOGITS

 excerpts

1.03

 affidav

1.01

 statements

0.98

 scathing

0.96

an

0.94

 detailed

0.93

0.91

 memos

0.91

 screenshots

0.89

 photographs

0.89

Activations Density 0.230%

news-related terms like "published" and "released."

instances of published or released information, typically from authoritative sources

No Comments

No Known Activations