INDEX

Explanations

statements of beliefs or opinions that support established views, as well as discussions on biases and accountability in belief systems

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

 Originally

-0.65

 Torrent

-0.63

 Walters

-0.58

undrum

-0.57

lda

-0.57

blogspot

-0.57

ircraft

-0.56

DragonMagazine

-0.55

outhern

-0.55

 Bennett

-0.54

POSITIVE LOGITS

)).

1.10

 theirs

0.99

 oneself

0.97

 themselves

0.91

?".

0.87

 others

0.86

.'"

0.81

them

0.81

'."

0.81

 injust

0.79

Activations Density 5.431%

statements of beliefs or opinions that support established views, as well as discussions on biases and accountability in belief systems

No Comments

No Known Activations