INDEX

Explanations

capital letters used for denoting individuals or titles

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

 envy

-0.70

Magikarp

-0.65

},{"

-0.64

 actionGroup

-0.63

olicy

-0.62

 encour

-0.62

 beforehand

-0.62

peak

-0.61

mods

-0.60

ulates

-0.60

POSITIVE LOGITS

UST

1.10

usterity

1.04

UTH

0.98

UG

0.95

 gunman

0.95

ircraft

0.94

cknowled

0.92

 woman

0.91

uckland

0.88

gency

0.86

Activations Density 0.062%

capital letters used for denoting individuals or titles

No Comments

No Known Activations