INDEX

Explanations

descriptions related to a new game or adventure

oai_token-act-pair · gpt-3.5-turbo

references to gameplay mechanics in video games

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

ļéĨĴ

-0.91

Fed

-0.77

 MSNBC

-0.77

 Oprah

-0.75

itutional

-0.75

ervative

-0.74

 Martha

-0.74

orgetown

-0.74

statement

-0.72

mercial

-0.72

POSITIVE LOGITS

 respawn

1.50

 NPCs

1.31

 gameplay

1.22

 playthrough

1.22

 teleport

1.22

 multiplayer

1.19

 melee

1.19

 quests

1.17

 turrets

1.13

 playable

1.11

Activations Density 0.804%

descriptions related to a new game or adventure

references to gameplay mechanics in video games

No Known Activations