INDEX

Explanations

supportive language and words related to advocating for a specific cause or belief

oai_token-act-pair · gpt-3.5-turbo

references to a specific cause or movement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

aeper

-0.78

ault

-0.75

 Leopard

-0.73

Ku

-0.69

 Pione

-0.69

olitan

-0.68

 Sheep

-0.68

lav

-0.66

 Centers

-0.66

 Technique

-0.65

POSITIVE LOGITS

 cele

1.27

cause

0.81

 Cause

0.79

way

0.76

 celeb

0.74

facts

0.73

wagon

0.71

forge

0.71

DNA

0.71

fare

0.70

Activations Density 0.028%

supportive language and words related to advocating for a specific cause or belief

references to a specific cause or movement

No Comments

No Known Activations