INDEX

Explanations

references to ways to provide feedback or support

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

Esk

-0.68

 Bord

-0.65

 Railroad

-0.64

 Twin

-0.64

 Exam

-0.63

 Scheme

-0.63

 Sacrament

-0.62

 Machines

-0.62

 Robot

-0.61

 Immortal

-0.61

POSITIVE LOGITS

1.18

pedia

0.97

archives

0.91

info

0.90

0.89

monkey

0.89

DonaldTrump

0.86

facts

0.85

oft

0.81

ickr

0.81

Activations Density 0.093%

references to ways to provide feedback or support

No Comments

No Known Activations