INDEX

Explanations

references to the word "Guardian"

oai_token-act-pair · gpt-3.5-turbo

mentions of the Guardian media outlet

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

merce

-0.85

itionally

-0.83

cles

-0.81

lease

-0.79

perm

-0.79

ramid

-0.77

anners

-0.73

zinski

-0.73

geon

-0.73

rano

-0.73

POSITIVE LOGITS

 Angels

0.99

 Angel

0.86

arians

0.84

ãĥķãĤ¡

0.83

 Islands

0.79

arium

0.78

 Agency

0.75

Leaks

0.74

 Editorial

0.74

arian

0.72

Activations Density 0.031%

references to the word "Guardian"

mentions of the Guardian media outlet

No Comments

No Known Activations