INDEX

Explanations

phrases related to government actions and policies

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

 myster

-0.72

 seiz

-0.70

 sacrific

-0.69

 federation

-0.68

 scattering

-0.67

 Negro

-0.64

 ende

-0.63

 notor

-0.62

 strugg

-0.62

 shadow

-0.62

POSITIVE LOGITS

ï¸ı

1.16

Â¯

1.05

cue

0.81

0.80

STEM

0.80

âĢł

0.78

Tea

0.77

Â°

0.76

endif

0.75

ï¸

0.74

Activations Density 0.382%

phrases related to government actions and policies

No Comments

No Known Activations