INDEX

Explanations

phrases related to statistical data and quantifiable information

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

ESE

-0.78

doors

-0.71

Interstitial

-0.70

emis

-0.69

CLA

-0.69

Activ

-0.68

ILE

-0.68

 Flavoring

-0.67

 Contents

-0.67

 Quit

-0.66

POSITIVE LOGITS

 inflated

1.15

 graphs

1.09

 underest

1.07

 skewed

1.05

 underestimate

1.02

 overest

1.00

 extrap

0.99

 statistical

0.98

 ballpark

0.98

 projections

0.98

Activations Density 0.200%

phrases related to statistical data and quantifiable information

No Comments

No Known Activations