INDEX

Explanations

terms related to stacking and structural components in programming or construction

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

amily

-0.68

 AMER

-0.63

 homegrown

-0.62

 domest

-0.60

manufact

-0.60

ERY

-0.58

 outfits

-0.57

 multinational

-0.57

 grassroots

-0.57

á½

-0.57

POSITIVE LOGITS

 respectively

0.94

 whence

0.83

endif

0.82

(=

0.80

stretched

0.77

 increments

0.76

 Activate

0.75

 versa

0.74

 THEN

0.74

Activations Density 0.177%

terms related to stacking and structural components in programming or construction

No Comments

No Known Activations