INDEX

Explanations

phrases or terms related to abstract concepts and complex social issues

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

orn

-0.71

 consolation

-0.68

 Finnish

-0.67

 Belfast

-0.63

 Byzantine

-0.62

 Georgian

-0.62

 constructive

-0.61

 Nordic

-0.61

 sewing

-0.61

 festive

-0.61

POSITIVE LOGITS

would

1.30

should

1.24

was

1.22

has

1.20

could

1.19

might

1.18

had

1.17

were

1.16

will

1.12

doesn

1.07

Activations Density 0.261%

phrases or terms related to abstract concepts and complex social issues

No Comments

No Known Activations