INDEX

Explanations

names and locations related to conflict or tension within a community

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

 Californ

-0.85

 Titanic

-0.83

 Chloe

-0.77

 FANTASY

-0.75

 Gravity

-0.75

 Uncharted

-0.74

 ponies

-0.72

 collision

-0.71

 champagne

-0.71

 conventions

-0.70

POSITIVE LOGITS

aq

1.29

awi

1.22

 Hasan

1.20

azi

1.18

jad

1.17

llah

1.16

rahim

1.14

hammad

1.12

 Ahmad

1.10

hari

1.10

Activations Density 0.179%

names and locations related to conflict or tension within a community

No Comments

No Known Activations