INDEX

Explanations

the word "just" followed by a numerical value indicating the strength of association

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

pora

-0.63

 Responsibility

-0.61

 Nurs

-0.57

 contention

-0.56

uable

-0.55

 Pett

-0.55

abal

-0.54

ahime

-0.54

closure

-0.53

aea

-0.53

POSITIVE LOGITS

ifications

1.15

ifiable

1.05

itia

0.91

IFIED

0.88

ices

0.84

if

0.84

ifier

0.84

IFIC

0.83

ifying

0.83

 kidding

0.82

Activations Density 0.061%

the word "just" followed by a numerical value indicating the strength of association

No Comments

No Known Activations