INDEX

Explanations

attributes related to in-game characters like attack damage, speed, and health points

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Kinn

-0.75

 Helsinki

-0.74

orld

-0.71

ovie

-0.69

Commercial

-0.64

 Homo

-0.64

sonian

-0.63

Klu

-0.62

 Telegram

-0.62

 Horizons

-0.62

POSITIVE LOGITS

 regeneration

1.16

 modifier

1.15

 cooldown

1.12

 modifiers

1.12

 bonuses

1.08

 potions

1.06

dmg

1.05

 bonus

1.02

 damage

0.99

 orbs

0.99

Activations Density 0.136%

attributes related to in-game characters like attack damage, speed, and health points

No Comments

No Known Activations