INDEX

Explanations

news articles and social media updates

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

pora

-0.68

 seiz

-0.64

 challeng

-0.63

 subsequ

-0.61

luster

-0.60

leneck

-0.57

 Archdemon

-0.56

tie

-0.56

der

-0.56

 necks

-0.56

POSITIVE LOGITS

ifiable

1.39

ifications

1.28

ified

0.97

if

0.96

ification

0.96

IFIED

0.94

itia

0.92

 kidding

0.92

ifi

0.90

ices

0.89

Activations Density 2.088%

news articles and social media updates

No Comments

No Known Activations