INDEX

Explanations

phrases related to anti-something, potentially indicating a focus on items or concepts associated with being against or counteracting something

oai_token-act-pair · gpt-3.5-turbo

terms related to anti-technology and anti-regulatory themes

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

 Halls

-0.76

shire

-0.72

 Chaff

-0.71

 Pelicans

-0.69

INGS

-0.67

ourcing

-0.67

 Remastered

-0.66

 Cosponsors

-0.66

 Kinn

-0.65

theless

-0.64

POSITIVE LOGITS

zac

0.89

war

0.87

hero

0.85

establishment

0.84

drug

0.84

strate

0.83

digy

0.81

hesis

0.81

abortion

0.81

anti

0.80

Activations Density 0.017%

phrases related to anti-something, potentially indicating a focus on items or concepts associated with being against or counteracting something

terms related to anti-technology and anti-regulatory themes

No Comments

No Known Activations