INDEX

Explanations

instances of recognizing or evaluating the quality of work or performance

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

Tang

-0.79

NetMessage

-0.75

fty

-0.71

ater

-0.66

raltar

-0.65

eros

-0.63

isks

-0.63

ities

-0.62

ories

-0.61

Phi

-0.61

POSITIVE LOGITS

 portraying

0.94

manship

0.91

 ethic

0.86

 illustrating

0.86

 capturing

0.83

 protecting

0.81

 showcasing

0.81

 combating

0.81

 keeping

0.80

 educating

0.79

Activations Density 0.016%

instances of recognizing or evaluating the quality of work or performance

No Comments

No Known Activations