INDEX

Explanations

statements related to political debates and policies

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

edia

-0.70

ery

-0.66

erick

-0.64

agram

-0.63

ETHOD

-0.62

etting

-0.61

ickers

-0.61

ridge

-0.60

rique

-0.60

icky

-0.59

POSITIVE LOGITS

except

1.61

 except

1.60

 Including

1.48

including

1.40

 including

1.39

 regardless

1.38

 irrespective

1.34

 INCLUD

1.25

excluding

1.17

 imaginable

1.16

Activations Density 0.424%

statements related to political debates and policies

No Comments

No Known Activations