INDEX

Explanations

technical terms related to software development and data analysis

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

 carbohyd

-0.74

ENN

-0.69

 citiz

-0.67

soDeliveryDate

-0.67

ãĥ¯ãĥ³

-0.67

ITNESS

-0.65

GoldMagikarp

-0.64

 Seym

-0.62

 antidepress

-0.61

 measurable

-0.60

POSITIVE LOGITS

hire

0.85

ttes

0.81

 extraord

0.78

âĦ¢

0.76

Â®

0.76

puff

0.71

urs

0.71

xual

0.70

tery

0.70

ails

0.69

Activations Density 5.901%

technical terms related to software development and data analysis

No Comments

No Known Activations