INDEX

Explanations

phrases related to racial disparities, discrimination, and social inequities

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

 drip

-0.73

amina

-0.71

 ellipt

-0.70

odium

-0.70

DCS

-0.68

ocket

-0.68

cold

-0.67

acles

-0.66

 Telesc

-0.66

pload

-0.65

POSITIVE LOGITS

 minorities

1.16

 supremacist

1.13

 slurs

1.12

 supremacists

1.12

ethnic

1.09

 backgrounds

1.01

 immigrants

1.01

 ethnicity

1.00

 males

1.00

 nationalist

1.00

Activations Density 4.396%

phrases related to racial disparities, discrimination, and social inequities

No Comments

No Known Activations