INDEX

Explanations

sections of text related to scientific research and data analysis

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 Disclosure

-0.59

 Replay

-0.59

 Shutterstock

-0.59

 Ragnar

-0.59

 Grateful

-0.58

 Stard

-0.57

 Encyclopedia

-0.56

 repertoire

-0.56

 embassies

-0.55

 opio

-0.55

POSITIVE LOGITS

 versa

1.09

 thereafter

0.96

 likewise

0.95

 thereof

0.86

itto

0.85

alion

0.80

ALSE

0.76

 thereto

0.74

oliath

0.74

aurus

0.74

Activations Density 5.222%

sections of text related to scientific research and data analysis

No Comments

No Known Activations