INDEX

Explanations

attributes related to sports performance, such as skills, abilities, and physical traits

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

 Funding

-0.85

 Journals

-0.84

 Governments

-0.82

 UNESCO

-0.81

 Editorial

-0.79

 commemor

-0.78

 Membership

-0.74

 Publication

-0.73

 Investig

-0.72

 Settlement

-0.72

POSITIVE LOGITS

 offensively

1.52

 defensively

1.44

 athleticism

1.32

 positional

1.29

 hitters

1.28

 midfield

1.28

 teammates

1.23

 athlet

1.19

 defensive

1.18

 tempo

1.18

Activations Density 2.374%

attributes related to sports performance, such as skills, abilities, and physical traits

No Comments

No Known Activations