INDEX

Explanations

information related to political figures, events, and controversies

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

xtap

-0.90

atan

-0.74

ĺħ

-0.68

IPS

-0.66

DragonMagazine

-0.65

uckle

-0.65

OCK

-0.65

uca

-0.65

Enlarge

-0.63

oes

-0.62

POSITIVE LOGITS

 than

1.54

 importantly

1.18

 Than

1.03

 stringent

0.99

than

0.97

ado

0.94

 sophisticated

0.92

 expensive

0.91

 nuanced

0.85

 complicated

0.85

Activations Density 0.738%

information related to political figures, events, and controversies

No Comments

No Known Activations