INDEX

Explanations

negations related to normal or positive situations

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

chuk

-0.66

 Seym

-0.62

enegger

-0.55

sylv

-0.54

Intern

-0.54

AMI

-0.53

{{

-0.52

hetti

-0.52

 Horizon

-0.52

aughtered

-0.52

POSITIVE LOGITS

ye

0.70

 hereby

0.61

actionDate

0.56

 largeDownload

0.55

bra

0.54

-)

0.53

 preferably

0.53

 congr

0.52

huh

0.51

':

0.51

Activations Density 0.362%

negations related to normal or positive situations

No Comments

No Known Activations