INDEX

Explanations

phrases related to specific science terminology such as "Planetary nebulae" and "exoplanets."

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

NetMessage

-0.77

dab

-0.69

 sensations

-0.64

 Kush

-0.61

 IMAGES

-0.61

 Flavoring

-0.60

 traff

-0.59

 vain

-0.59

govtrack

-0.59

Downloadha

-0.58

POSITIVE LOGITS

etary

1.30

escape

1.09

isphere

1.07

ning

1.03

emaker

0.98

Ã©t

0.98

ets

0.97

ters

0.95

iverse

0.93

acci

0.90

Activations Density 0.025%

phrases related to specific science terminology such as "Planetary nebulae" and "exoplanets."

No Comments

No Known Activations