INDEX

Explanations

phrases indicating different sides of an argument

oai_token-act-pair · gpt-3.5-turbo

instances of the word "argue" and its variations, highlighting discussions or debates

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Portable

-0.61

grad

-0.60

 Grad

-0.58

ha

-0.57

 Perfect

-0.57

 Powered

-0.56

 direction

-0.56

ent

-0.55

 cess

-0.55

Alert

-0.55

POSITIVE LOGITS

 argue

3.45

 argued

2.14

 argues

2.10

 arguing

2.03

 contend

1.87

 accuse

1.57

 argument

1.53

 speculate

1.52

 contends

1.50

 arguments

1.45

Activations Density 0.012%

phrases indicating different sides of an argument

instances of the word "argue" and its variations, highlighting discussions or debates

No Comments

No Known Activations