INDEX

Explanations

phrases related to joking or humor

oai_token-act-pair · gpt-3.5-turbo

instances of humor and speculation in the text

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

waters

-0.73

edom

-0.71

enf

-0.69

ãĥ¡

-0.69

BO

-0.68

por

-0.68

ĪĴ

-0.66

water

-0.66

orneys

-0.62

ebook

-0.61

POSITIVE LOGITS

 sarcast

0.81

 omin

0.79

 aloud

0.73

llor

0.67

 angrily

0.65

 gloom

0.63

 mourn

0.63

 Wilde

0.63

aults

0.63

atively

0.63

Activations Density 0.052%

phrases related to joking or humor

instances of humor and speculation in the text

No Comments

No Known Activations