INDEX

Explanations

words related to controversial topics or debates

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

 Kubrick

-0.72

 Refuge

-0.65

 Icelandic

-0.63

 Labyrinth

-0.62

 inund

-0.62

 gratification

-0.61

 circle

-0.60

 Ducks

-0.59

 Desire

-0.56

 repetitive

-0.56

POSITIVE LOGITS

ptroller

1.27

fortable

1.26

puters

1.25

pleting

1.22

pletion

1.21

forts

1.18

mented

1.18

puting

1.10

ittee

1.09

parable

1.09

Activations Density 0.539%

words related to controversial topics or debates

No Comments

No Known Activations