INDEX

Explanations

references to related topics or articles

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

stood

-0.75

raf

-0.74

-0.71

phia

-0.70

atur

-0.68

amping

-0.65

adish

-0.63

 recl

-0.63

arer

-0.63

angers

-0.62

POSITIVE LOGITS

 STOR

1.23

 VIDEOS

1.21

 INFORMATION

1.10

 STORY

1.08

 ARTICLE

1.04

RELATED

1.01

 VIDEO

0.99

 REPORT

0.95

 APPLIC

0.94

 UPDATE

0.93

Activations Density 0.004%

references to related topics or articles

No Comments

No Known Activations