INDEX

Explanations

mentions of the news channel "Fox News"

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 @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

acterial

-0.77

rians

-0.75

akeru

-0.71

rian

-0.71

 inval

-0.70

attled

-0.69

ansom

-0.68

ochemical

-0.68

Definition

-0.67

orically

-0.67

POSITIVE LOGITS

conn

1.36

hawk

0.92

Fox

0.88

 News

0.84

Fox

0.83

fox

0.83

woods

0.82

cat

0.79

News

0.78

FOX

0.77

Activations Density 0.015%

mentions of the news channel "Fox News"

references to Fox News

No Comments

No Known Activations