INDEX

Explanations

numerical data or lists

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

 fabricated

-0.66

ascript

-0.65

facade

-0.62

 sampling

-0.62

 Parables

-0.61

 scare

-0.61

 plaque

-0.61

vertisement

-0.60

PIN

-0.59

POSITIVE LOGITS

etc

0.98

 TOTAL

0.84

Others

0.78

 Flavoring

0.77

 Lastly

0.73

 Finally

0.72

 Conclusion

0.71

Conclusion

0.71

 âĶľ

0.70

rador

0.69

Activations Density 0.157%

numerical data or lists

No Comments

No Known Activations