INDEX

Explanations

references to political events and processes, such as referendums and political 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_fs6144-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_6144

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

6,144

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

etter

-0.73

Kut

-0.73

iak

-0.72

esthes

-0.70

iewicz

-0.68

paces

-0.68

 tremend

-0.68

ienne

-0.65

ATA

-0.65

ologies

-0.65

POSITIVE LOGITS

 referendum

1.21

ndum

1.13

 vote

0.93

 petition

0.87

 refere

0.86

ignty

0.82

 voting

0.81

 votes

0.80

 ballot

0.78

 polling

0.74

Activations Density 0.413%

references to political events and processes, such as referendums and political debates

No Comments

No Known Activations