INDEX

Explanations

advertisements in a document

oai_token-act-pair · gpt-3.5-turbo

indicators of advertisement or promotional content

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

ĪĴ

-0.76

opian

-0.72

rency

-0.72

ħĭ

-0.67

acea

-0.66

 ashes

-0.63

 fruitful

-0.63

uality

-0.62

amiya

-0.62

rite

-0.62

POSITIVE LOGITS

][/

0.76

 WATCHED

0.74

 sidx

0.68

eh

0.66

Jac

0.66

]"

0.66

advertisement

0.64

inately

0.63

 Abrams

0.62

 IMAGES

0.61

Activations Density 0.071%

advertisements in a document

indicators of advertisement or promotional content

No Comments

No Known Activations