INDEX

Explanations

mentions of the Fox News network

oai_token-act-pair · gpt-3.5-turbo

references to "Fox News."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 inval

-0.64

quo

-0.64

orically

-0.62

adding

-0.61

rians

-0.61

ussed

-0.60

umbn

-0.60

 captcha

-0.60

accompan

-0.59

iments

-0.59

POSITIVE LOGITS

conn

1.53

 News

1.06

News

1.02

croft

0.95

woods

0.93

borough

0.92

 Broadcasting

0.91

cat

0.88

Sports

0.87

xy

0.86

Activations Density 0.021%

mentions of the Fox News network

references to "Fox News."

No Comments

No Known Activations