INDEX

Explanations

phrases related to negation or contrast

oai_token-act-pair · gpt-3.5-turbo

negative contractions and phrases related to denial or refusal

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

furt

-0.87

ĨĴ

-0.76

soever

-0.73

bur

-0.72

ipers

-0.71

Ã»

-0.67

orie

-0.67

NCT

-0.67

wich

-0.65

ãĥİ

-0.65

POSITIVE LOGITS

 already

0.78

hin

0.66

EEK

0.65

 occasionally

0.62

MAR

0.60

grown

0.59

 plenty

0.59

anan

0.58

abo

0.58

lamm

0.56

Activations Density 0.210%

phrases related to negation or contrast

negative contractions and phrases related to denial or refusal

No Comments

No Known Activations