INDEX

Explanations

phrases related to opinions and beliefs

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 pione

-0.73

pez

-0.71

gur

-0.70

Vers

-0.66

yna

-0.65

haul

-0.65

PokÃ©

-0.62

ufact

-0.62

ultz

-0.61

iets

-0.61

POSITIVE LOGITS

 anybody

1.23

 anyone

1.15

 anymore

1.07

 anything

1.07

any

0.99

 whatsoever

0.97

nor

0.96

 anywhere

0.84

ANY

0.81

 EVER

0.81

Activations Density 0.045%

phrases related to opinions and beliefs

No Comments

No Known Activations