INDEX

Explanations

the word "mark" followed by a number

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

ILLE

-0.90

anooga

-0.77

 traged

-0.69

rouch

-0.68

 brut

-0.67

ettel

-0.67

åŃ

-0.66

abama

-0.66

aband

-0.65

ocal

-0.64

POSITIVE LOGITS

manship

1.49

downs

1.11

down

0.99

emark

0.92

 Twain

0.90

ups

0.89

posts

0.88

eters

0.86

marks

0.84

lights

0.83

Activations Density 0.022%

the word "mark" followed by a number

No Comments

No Known Activations