INDEX

Explanations

terms related to programming and programming languages

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

-0.82

urn

-0.80

ously

-0.78

ewitness

-0.78

ous

-0.77

lain

-0.75

vol

-0.75

yson

-0.75

ividual

-0.74

iful

-0.68

POSITIVE LOGITS

 Languages

1.22

 languages

1.20

mers

1.00

matically

0.99

mable

0.94

 interfaces

0.92

 language

0.90

Blocks

0.86

matic

0.85

 theorem

0.82

Activations Density 0.010%

terms related to programming and programming languages

No Comments

No Known Activations