INDEX

Explanations

special characters or symbols that often indicate formatting or coding elements

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

 subsequ

-0.87

 confir

-0.86

 tremend

-0.86

 sugg

-0.82

ĺħ

-0.82

 livest

-0.79

 conduc

-0.77

 probation

-0.77

 streng

-0.75

regon

-0.74

POSITIVE LOGITS

////////////////

1.30

////////

1.24

////////////////////////////////

1.21

wcsstore

0.97

////

0.96

++++++++++++++++

0.92

SIGN

0.82

pad

0.80

www

0.76

ECH

0.75

Activations Density 0.004%

special characters or symbols that often indicate formatting or coding elements

No Comments

No Known Activations