INDEX

Explanations

references to politics and controversies, especially related to specific events or individuals

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

 exception

-0.63

 Portuguese

-0.60

ASED

-0.58

 citation

-0.57

 adjunct

-0.57

MENTS

-0.57

 submission

-0.57

 reinforcement

-0.57

aneously

-0.56

 Sweeney

-0.56

POSITIVE LOGITS

mith

1.78

hift

1.69

pace

1.66

ilver

1.65

chool

1.61

creen

1.61

peed

1.60

hip

1.60

omething

1.58

pring

1.57

Activations Density 2.036%

references to politics and controversies, especially related to specific events or individuals

No Comments

No Known Activations