INDEX

Explanations

terms related to law, crime, justice, and legal procedures

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

arity

-0.73

rish

-0.70

 Pixie

-0.70

metry

-0.67

blank

-0.67

por

-0.67

mate

-0.66

galitarian

-0.64

pole

-0.63

mates

-0.63

POSITIVE LOGITS

 punishable

1.02

 spree

1.00

 mastermind

0.98

 convictions

0.96

 offenses

0.95

ously

0.93

 prosecutions

0.92

 offences

0.91

 gangs

0.90

ous

0.89

Activations Density 0.744%

terms related to law, crime, justice, and legal procedures

No Comments

No Known Activations