INDEX

Explanations

bold actions or statements

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

OTOS

-0.90

 Cheong

-0.87

enfranch

-0.65

 duly

-0.64

ADS

-0.63

utra

-0.63

AW

-0.62

yip

-0.62

apolis

-0.62

nesota

-0.61

POSITIVE LOGITS

faced

1.20

er

1.09

ness

1.05

face

0.98

est

0.89

mouth

0.89

nesses

0.84

ly

0.82

bold

0.81

word

0.81

Activations Density 0.028%

bold actions or statements

No Comments

No Known Activations