INDEX

Explanations

individual names or titles

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

creen

-0.70

 Mandela

-0.68

grass

-0.66

SOS

-0.63

 Schumer

-0.62

 Corpus

-0.62

terson

-0.62

rium

-0.61

 Libre

-0.59

 resume

-0.58

POSITIVE LOGITS

senal

1.39

haps

1.26

outine

1.24

ansom

1.23

ascal

1.17

itchie

1.17

ussia

1.13

abbit

1.08

ework

1.08

ifles

1.08

Activations Density 2.003%

individual names or titles

No Comments

No Known Activations