INDEX

Explanations

instances of the word "persons"

oai_token-act-pair · gpt-3.5-turbo

references to individuals or groups of people

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

Accessory

-0.82

Phill

-0.73

enthal

-0.70

ļéĨĴ

-0.69

 recess

-0.67

 Peck

-0.67

è¡

-0.66

âĺħâĺħ

-0.65

NES

-0.64

 Ridge

-0.64

POSITIVE LOGITS

 persons

0.95

folk

0.91

hood

0.89

soever

0.79

isms

0.79

cript

0.79

metics

0.76

 Persons

0.75

 blinded

0.74

acters

0.74

Activations Density 0.011%

instances of the word "persons"

references to individuals or groups of people

No Comments

No Known Activations