INDEX

Explanations

instances of self-reflection and personal insights

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

Gad

-0.66

 antitrust

-0.62

 Canaver

-0.60

icist

-0.60

 optics

-0.58

 extant

-0.58

 Rockefeller

-0.57

 Plaint

-0.57

 Millennium

-0.57

 Consortium

-0.56

POSITIVE LOGITS

'm

1.65

've

1.35

'll

1.22

am

1.15

'd

1.12

 guess

1.06

owe

1.02

 adore

1.02

 swear

1.02

don

1.01

Activations Density 0.266%

instances of self-reflection and personal insights

No Comments

No Known Activations