INDEX

Explanations

words related to appropriation and misappropriation

oai_token-act-pair · gpt-3.5-turbo

instances and discussions of cultural appropriation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Bay

-0.69

lock

-0.69

Sau

-0.69

Box

-0.69

Oracle

-0.68

 sail

-0.68

 Solitaire

-0.68

 Sussex

-0.68

dream

-0.68

 Holmes

-0.67

POSITIVE LOGITS

appropri

1.45

 appropriation

1.14

ropri

1.11

 appropri

1.10

 appropriated

1.04

appropriately

1.00

itures

0.93

appropriate

0.91

orrow

0.88

anity

0.86

Activations Density 0.019%

words related to appropriation and misappropriation

instances and discussions of cultural appropriation

No Comments

No Known Activations