INDEX

Explanations

phrases related to narratives about events or actions being done by multiple individuals

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

Negative Logits

itaire

-0.57

icter

-0.56

 Ambro

-0.55

 advoc

-0.49

acerb

-0.49

asant

-0.47

Basic

-0.47

earcher

-0.46

asus

-0.46

pex

-0.46

POSITIVE LOGITS

 themselves

1.04

 selves

0.94

selves

0.92

 asses

0.64

 mouths

0.63

MpServer

0.62

 necks

0.62

 their

0.61

 yourselves

0.61

 souls

0.60

Activations Density 0.928%

phrases related to narratives about events or actions being done by multiple individuals

No Known Activations