INDEX

Explanations

words related to morality and judgment, especially focusing on what is considered offensive or harmful

oai_token-act-pair · gpt-3.5-turbo

terms related to offensive or harmful content and their effects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

onut

-0.75

riage

-0.66

prise

-0.63

abet

-0.61

aper

-0.60

issue

-0.59

icans

-0.59

anche

-0.58

avis

-0.56

cknowled

-0.56

POSITIVE LOGITS

 enough

1.04

enough

0.89

 outwe

0.87

 Enough

0.84

 beforehand

0.75

terday

0.74

or

0.73

 outweigh

0.73

.''

0.70

Enough

0.65

Activations Density 0.388%

words related to morality and judgment, especially focusing on what is considered offensive or harmful

terms related to offensive or harmful content and their effects

No Comments

No Known Activations