INDEX

Explanations

phrases related to law enforcement leadership or high-ranking police officials

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

nikov

-0.78

ted

-0.78

schild

-0.75

pmwiki

-0.75

mented

-0.73

Sov

-0.73

\/\/

-0.73

adesh

-0.73

AAF

-0.70

dates

-0.70

POSITIVE LOGITS

doms

1.05

 Petty

0.91

IAL

0.81

pin

0.80

Chief

0.74

 Warrant

0.74

 Darrell

0.71

 Marshal

0.71

Jay

0.68

 Kelvin

0.68

Activations Density 0.018%

phrases related to law enforcement leadership or high-ranking police officials

No Comments

No Known Activations