INDEX

Explanations

references to social issues and political topics

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

 mathemat

-0.76

 sacrific

-0.72

 fortun

-0.68

 loopholes

-0.66

 scattering

-0.64

 elig

-0.64

 myster

-0.64

 elim

-0.62

jog

-0.61

 SERV

-0.61

POSITIVE LOGITS

ï¸ı

1.39

ski

0.89

mental

0.86

tracks

0.86

0.84

sure

0.82

ttle

0.81

esc

0.81

ship

0.80

ï¸

0.80

Activations Density 0.772%

references to social issues and political topics

No Comments

No Known Activations