INDEX

Explanations

proper names and terms in German

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

theless

-0.76

 compulsion

-0.69

 captcha

-0.68

 shockingly

-0.64

 VIDE

-0.63

 bullies

-0.62

 microw

-0.61

 duplication

-0.60

inarily

-0.60

 bonded

-0.60

POSITIVE LOGITS

liga

0.97

 MÃ©

0.90

 Ãī

0.88

qui

0.86

 Univers

0.85

usalem

0.85

arten

0.81

å£«

0.81

Ãī

0.81

aceae

0.79

Activations Density 0.154%

proper names and terms in German

No Comments

No Known Activations