INDEX

Explanations

words related to legal or criminal activities

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

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

PRES

-0.76

hyde

-0.74

âĸ¬

-0.74

Madison

-0.73

gerald

-0.71

NEY

-0.70

lish

-0.68

BOOK

-0.67

LY

-0.67

SPONSORED

-0.67

POSITIVE LOGITS

atter

1.26

aques

1.23

umber

1.22

umbers

1.20

acer

1.15

iers

1.14

umb

1.12

acent

1.10

ump

1.08

asma

1.06

Activations Density 0.014%

words related to legal or criminal activities

No Comments

No Known Activations