INDEX

Explanations

words related to drug use and illegal 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

xual

-0.72

aeda

-0.72

heid

-0.66

EMENT

-0.64

mentation

-0.64

 Centauri

-0.63

Constructed

-0.62

ments

-0.61

BAT

-0.60

upon

-0.60

POSITIVE LOGITS

ety

1.11

okin

1.10

atches

1.02

appy

1.02

ipes

0.99

igger

0.99

ickers

0.99

eties

0.98

uggle

0.97

atching

0.95

Activations Density 0.836%

words related to drug use and illegal activities

No Comments

No Known Activations