INDEX

Explanations

terms related to crime and 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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

lihood

-0.81

ENC

-0.73

 Polo

-0.72

EMENT

-0.69

é¾įå

-0.68

IELD

-0.68

ku

-0.67

 Flickr

-0.67

nah

-0.66

ï¸

-0.64

POSITIVE LOGITS

umb

0.94

onut

0.93

atism

0.91

umbs

0.87

umbing

0.82

adle

0.82

ursor

0.81

umbles

0.80

anium

0.79

utch

0.78

Activations Density 0.770%

terms related to crime and criminal activities

No Comments

No Known Activations