INDEX

Explanations

references to law enforcement, police activities, and legal policies

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 Gree

-0.67

 Scion

-0.65

 Wast

-0.63

 Expedition

-0.62

 Cros

-0.61

 Mous

-0.61

Dak

-0.60

hiba

-0.59

 Flavoring

-0.58

 Vice

-0.58

POSITIVE LOGITS

 enforcement

1.37

enforcement

1.29

breakers

1.24

makers

1.23

breaker

1.20

lessness

1.19

abiding

1.19

breaking

1.16

 Enforcement

1.15

 abiding

1.14

Activations Density 1.551%

references to law enforcement, police activities, and legal policies

No Comments

No Known Activations