INDEX

Explanations

phrases related to providing instructions or explanations

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

instances of possessive constructions, indicating ownership or association

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

soever

-0.66

roit

-0.63

estic

-0.62

onte

-0.62

ottesville

-0.60

okers

-0.59

ellig

-0.57

æĪ¦

-0.57

doms

-0.57

lees

-0.57

POSITIVE LOGITS

 hoping

1.11

how

1.07

why

1.06

 what

1.02

 whats

0.89

 where

0.86

 another

0.83

my

0.79

WHY

0.78

HOW

0.77

Activations Density 0.076%

phrases related to providing instructions or explanations

instances of possessive constructions, indicating ownership or association

No Comments

No Known Activations