INDEX

Explanations

phrases related to political and social issues

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

 shape

-0.80

 habit

-0.78

 carrier

-0.76

 salv

-0.73

 barg

-0.73

 downs

-0.72

 likeness

-0.72

 pharmacy

-0.72

 nude

-0.72

 disguise

-0.71

POSITIVE LOGITS

ï¸ı

1.19

ï¸

1.06

 Similarly

1.02

 Additionally

1.02

 Unfortunately

1.01

However

0.99

 Therefore

0.98

 However

0.98

 Likewise

0.95

 Obviously

0.94

Activations Density 0.465%

phrases related to political and social issues

No Comments

No Known Activations