INDEX

Explanations

names and entities related to prominent public figures or events

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

ortium

-0.68

nces

-0.66

natureconservancy

-0.64

yip

-0.64

abase

-0.62

perture

-0.61

ragon

-0.60

 guiName

-0.59

href

-0.58

rongh

-0.57

POSITIVE LOGITS

schild

0.78

 Sabha

0.66

 Janeiro

0.64

onding

0.62

ooth

0.62

runner

0.60

owsky

0.59

obin

0.59

riage

0.58

owitz

0.58

Activations Density 5.312%

names and entities related to prominent public figures or events

No Comments

No Known Activations