INDEX

Explanations

specific word endings, particularly those that denote relationships or actions

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

yip

-0.64

lington

-0.57

 Invalid

-0.57

 Carney

-0.56

 Camer

-0.56

idon

-0.56

 Shutterstock

-0.55

 Falk

-0.55

 Parish

-0.55

cair

-0.55

POSITIVE LOGITS

pees

0.60

opoly

0.58

Ã®

0.55

ogi

0.55

 marble

0.54

gger

0.54

irs

0.54

 Ãł

0.52

raq

0.52

uce

0.52

Activations Density 0.054%

specific word endings, particularly those that denote relationships or actions

No Comments

No Known Activations