INDEX

Explanations

references to popular fictional characters and related terms, specifically "Zelda" and "Korra."

oai_token-act-pair · gpt-3.5-turbo

references to specific video game characters and franchises

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 captcha

-0.64

rikes

-0.63

lain

-0.62

 bombard

-0.61

ynski

-0.59

 glim

-0.59

lv

-0.58

ripp

-0.58

 warrants

-0.57

lying

-0.57

POSITIVE LOGITS

lda

1.12

 Zelda

1.12

meric

0.81

weed

0.81

 Breath

0.76

 Korra

0.74

Wii

0.74

Pg

0.72

Å

0.72

 Metroid

0.70

Activations Density 0.014%

references to popular fictional characters and related terms, specifically "Zelda" and "Korra."

references to specific video game characters and franchises

No Comments

No Known Activations