INDEX

Explanations

phrases associated with violent or conflict-related events or contexts

oai_token-act-pair · gpt-3.5-turbo

words related to identities and categories

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

enegger

-0.84

elf

-0.82

raised

-0.79

orate

-0.73

oken

-0.71

raising

-0.68

CHA

-0.67

irth

-0.64

ioned

-0.63

 Broken

-0.62

POSITIVE LOGITS

idal

1.43

 pend

0.76

ãĥ³ãĤ¸

0.69

ãĥ¥

0.69

ity

0.67

ysis

0.67

atory

0.66

oad

0.66

itous

0.65

ITIES

0.64

Activations Density 0.010%

phrases associated with violent or conflict-related events or contexts

words related to identities and categories

No Comments

No Known Activations