INDEX

Explanations

mentions of gunfire

oai_token-act-pair · gpt-3.5-turbo

references to gunfire and gun-related incidents

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Label

-0.83

orse

-0.80

Wallet

-0.79

Stud

-0.79

val

-0.77

vals

-0.73

obar

-0.71

Sat

-0.70

ebook

-0.70

BD

-0.69

POSITIVE LOGITS

 gunfire

1.20

 bullets

1.16

 gunshots

1.02

 barr

1.01

 ammunition

0.95

 Bullets

0.92

 ammo

0.87

anguage

0.83

 shooting

0.83

 projectiles

0.77

Activations Density 0.015%

mentions of gunfire

references to gunfire and gun-related incidents

No Comments

No Known Activations