INDEX

Explanations

terms related to the concept of "wholeness" or "collective."

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

minist

-0.83

ufact

-0.81

 Machines

-0.79

 Nurs

-0.67

Fax

-0.65

 Canter

-0.65

 Specialist

-0.64

 Sunder

-0.63

 Cert

-0.62

 Lawyers

-0.62

POSITIVE LOGITS

heartedly

1.09

hearted

0.94

meal

0.76

 comprised

0.65

,...

0.64

 ensemble

0.64

 comprising

0.63

 standpoint

0.60

owe

0.60

 length

0.60

Activations Density 0.012%

terms related to the concept of "wholeness" or "collective."

No Comments

No Known Activations