INDEX

Explanations

references to the Fox News network

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "Fox" as it relates to news media

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rians

-0.75

 inval

-0.72

acterial

-0.71

assy

-0.70

akeru

-0.69

rian

-0.69

rogens

-0.69

ansom

-0.67

rella

-0.65

ictions

-0.65

POSITIVE LOGITS

conn

1.53

 News

0.94

News

0.90

cat

0.88

woods

0.87

croft

0.86

hawk

0.85

es

0.81

fire

0.81

Wire

0.80

Activations Density 0.022%

references to the Fox News network

occurrences of the term "Fox" as it relates to news media

No Comments

No Known Activations