INDEX

Explanations

terms related to religious institutions, particularly the word "temple"

oai_token-act-pair · gpt-3.5-turbo

references to temples

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

Buzz

-0.83

 Canad

-0.83

 Buzz

-0.81

iability

-0.72

Analysis

-0.71

 McInt

-0.70

 Crosby

-0.68

 Loaded

-0.68

 Holl

-0.67

 Stras

-0.66

POSITIVE LOGITS

 temple

3.78

 temples

3.32

 Temple

2.20

 shrine

1.86

 altar

1.64

 monastery

1.58

 synagogue

1.58

 mosque

1.52

 chapel

1.50

 priesthood

1.41

Activations Density 0.009%

terms related to religious institutions, particularly the word "temple"

references to temples

No Comments

No Known Activations