INDEX

Explanations

references to figures or illustrations in a text

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

ngth

-0.79

UGH

-0.68

 Vaugh

-0.67

 Sussex

-0.66

 administ

-0.66

 regist

-0.65

×ĳ

-0.63

 Passenger

-0.62

TPPStreamerBot

-0.62

 Sacrament

-0.62

POSITIVE LOGITS

ured

1.23

uring

1.21

ures

1.20

uration

1.20

aro

1.08

uer

1.05

ue

1.03

uers

1.02

urations

1.01

ues

0.95

Activations Density 0.003%

references to figures or illustrations in a text

No Comments

No Known Activations