INDEX

Explanations

words related to applying for or being involved in various applications, grants, or programs

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

ifa

-0.73

hov

-0.73

footed

-0.68

airs

-0.66

esses

-0.64

fall

-0.64

tery

-0.63

ewater

-0.63

ottesville

-0.63

heimer

-0.62

POSITIVE LOGITS

 pressure

0.68

 Applic

0.68

 mathemat

0.68

 deadline

0.67

atana

0.67

 scrut

0.67

 reluct

0.66

atism

0.66

 appl

0.65

arton

0.65

Activations Density 0.599%

words related to applying for or being involved in various applications, grants, or programs

No Comments

No Known Activations