INDEX

Explanations

proper nouns, such as names of people and places

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate cause and effect relationships

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

csv

-0.85

md

-0.78

izons

-0.77

pages

-0.71

each

-0.69

oother

-0.68

mins

-0.67

existent

-0.65

 Oliv

-0.65

ocket

-0.64

POSITIVE LOGITS

 homage

1.44

 ploy

1.31

 attempt

1.29

 indication

1.26

 reflection

1.24

 tribute

1.24

 manifestation

1.23

 acknowledgment

1.20

 consequence

1.20

 retaliation

1.17

Activations Density 0.272%

proper nouns, such as names of people and places

phrases that indicate cause and effect relationships

No Comments

No Known Activations