INDEX

Explanations

news-related terms, specifically mentioning law enforcement and criminal activities

oai_token-act-pair · gpt-3.5-turbo

mentions of law enforcement entities and their activities

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

chery

-0.77

vote

-0.75

born

-0.74

perm

-0.73

arist

-0.72

opian

-0.71

oscope

-0.70

itary

-0.68

cast

-0.67

thur

-0.66

POSITIVE LOGITS

 Constable

0.95

 Commissioner

0.86

 Dept

0.83

 Enforcement

0.77

 Files

0.76

 Reporter

0.75

 Officers

0.75

 Officer

0.74

 Superintendent

0.74

 Cust

0.74

Activations Density 0.056%

news-related terms, specifically mentioning law enforcement and criminal activities

mentions of law enforcement entities and their activities

No Comments

No Known Activations

news-related terms, specifically mentioning law enforcement and criminal activities

mentions of law enforcement entities and their activities

No Comments

No Known Activations