INDEX

Explanations

mentions of the word "Che" with high activation values

oai_token-act-pair · gpt-3.5-turbo

references to a specific character or entity named "Che."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

SI

-0.69

 reprint

-0.68

 maternity

-0.66

ML

-0.65

 tornado

-0.65

 lighthouse

-0.65

 DACA

-0.64

 subdivision

-0.63

MLA

-0.62

ER

-0.62

POSITIVE LOGITS

Che

4.00

Che

2.13

che

2.07

Chel

1.69

che

1.52

cheat

1.25

 Cheng

1.19

 Cheese

1.18

Chicken

1.17

Chem

1.16

Activations Density 0.017%

mentions of the word "Che" with high activation values

references to a specific character or entity named "Che."

No Comments

No Known Activations