INDEX

Explanations

news sources and media-related words

oai_token-act-pair · gpt-3.5-turbo

mentions of news organizations and their broadcasts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

naire

-0.76

tee

-0.71

sed

-0.71

Wah

-0.67

 FedEx

-0.67

 Dropbox

-0.66

ouses

-0.65

 groom

-0.64

 Haku

-0.64

 hypert

-0.62

POSITIVE LOGITS

radio

0.84

 Correspond

0.74

iland

0.72

 Wales

0.70

liv

0.70

isbury

0.69

levision

0.69

Ãį

0.69

INESS

0.68

 Insider

0.67

Activations Density 0.139%

news sources and media-related words

mentions of news organizations and their broadcasts

No Comments

No Known Activations