INDEX

Explanations

terms related to the importance or significance of specific factors, especially in the context of medical, scientific, and societal discussions

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

renheit

-0.83

uthor

-0.79

©¶æ

-0.77

 Fever

-0.76

 Carbuncle

-0.75

ibal

-0.74

fred

-0.73

çİĭ

-0.73

bows

-0.71

ilk

-0.70

POSITIVE LOGITS

 ingredient

1.00

 importance

0.98

stone

0.90

 distinction

0.85

 ingred

0.85

 component

0.82

 components

0.81

 hinge

0.79

 aspect

0.78

 aspects

0.78

Activations Density 1.571%

terms related to the importance or significance of specific factors, especially in the context of medical, scientific, and societal discussions

No Comments

No Known Activations