INDEX

Explanations

proper nouns related to media and politics

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

Not in Any Lists

No Comments

Negative Logits

 FANTASY

-0.67

 Confederation

-0.67

ruary

-0.67

ModLoader

-0.66

å§«

-0.65

 depend

-0.60

 Sioux

-0.60

Ary

-0.59

 bitters

-0.59

 Leia

-0.59

POSITIVE LOGITS

isner

0.74

uner

0.72

usb

0.67

acher

0.67

onson

0.67

roth

0.67

ãĤ±

0.66

itzer

0.64

adiq

0.64

opol

0.63

Activations Density 0.067%

proper nouns related to media and politics

No Comments

No Known Activations