INDEX

Explanations

information related to specific news or events, potentially related to court cases or investigations

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_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

abol

-0.70

planet

-0.69

poon

-0.68

dict

-0.68

scientific

-0.66

gross

-0.64

Redd

-0.64

rh

-0.63

soType

-0.63

ring

-0.63

POSITIVE LOGITS

 IMAGES

1.15

 FILE

1.13

 PHOTO

1.13

 Protesters

1.03

 VIDEO

0.90

allery

0.80

 Newly

0.79

 Aerial

0.79

 Scenes

0.79

 PRESS

0.78

Activations Density 0.034%

information related to specific news or events, potentially related to court cases or investigations

No Comments

No Known Activations