INDEX

Explanations

references to protests and demonstrations, particularly related to social and political activism

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

fixed

-0.72

neau

-0.72

 curv

-0.71

washer

-0.70

 probabilities

-0.68

availability

-0.67

liest

-0.66

Ö

-0.65

 Acqu

-0.64

puted

-0.63

POSITIVE LOGITS

 protesting

1.32

 protest

1.30

 activists

1.23

 Protest

1.21

 protestors

1.18

 protests

1.14

 protesters

1.13

 denouncing

1.13

 activist

1.13

 rallying

1.13

Activations Density 6.550%

references to protests and demonstrations, particularly related to social and political activism

No Comments

No Known Activations