INDEX

Explanations

phrases related to global or worldwide entities or concepts

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

ĪĴ

-0.84

ADRA

-0.77

¬¼

-0.74

regon

-0.73

hift

-0.73

zes

-0.72

POR

-0.71

zin

-0.71

okane

-0.71

irc

-0.69

POSITIVE LOGITS

(~

0.74

 besides

0.74

 according

0.71

 today

0.69

 compared

0.67

 âĶ

0.66

 bestowed

0.65

 excluding

0.64

 surpassed

0.64

 EVER

0.63

Activations Density 0.112%

phrases related to global or worldwide entities or concepts

No Comments

No Known Activations