INDEX

Explanations

lines of code including loops and indexing through arrays

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

andre

-0.70

rote

-0.69

ibur

-0.64

âĵĺ

-0.62

ãĥĦ

-0.61

ptin

-0.60

 largeDownload

-0.59

AIR

-0.56

FX

-0.56

FontSize

-0.56

POSITIVE LOGITS

!=

0.71

NUM

0.67

 enumer

0.66

 Length

0.64

){

0.63

ipher

0.63

range

0.63

è¦ļéĨĴ

0.63

 âī¤

0.63

vals

0.62

Activations Density 0.042%

lines of code including loops and indexing through arrays

No Comments

No Known Activations