INDEX

Explanations

the word "chose"

oai_token-act-pair · gpt-3.5-turbo

instances of the word "chose" and its variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

emic

-0.75

aging

-0.75

cre

-0.72

pat

-0.71

uum

-0.67

otor

-0.67

Jam

-0.67

oidal

-0.65

ear

-0.65

ijk

-0.65

POSITIVE LOGITS

 chose

0.99

 chooses

0.95

 choose

0.87

ļéĨĴ

0.85

 chosen

0.84

èĢħ

0.79

($)

0.76

 choices

0.75

 opted

0.75

 Compass

0.74

Activations Density 0.013%

the word "chose"

instances of the word "chose" and its variations

No Comments

No Known Activations