INDEX

Explanations

instances where the word "might" is used

oai_token-act-pair · gpt-3.5-turbo

instances of the word "might"

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

oric

-0.89

cake

-0.67

 VIDEOS

-0.66

 Purpose

-0.65

scar

-0.65

zen

-0.64

carbon

-0.64

LOT

-0.63

ross

-0.62

ICO

-0.62

POSITIVE LOGITS

 feas

0.97

 plaus

0.96

 tremend

0.89

 sugg

0.86

berra

0.85

 conce

0.84

 conclud

0.81

 theoretically

0.80

 mistakenly

0.79

haps

0.79

Activations Density 0.029%

instances where the word "might" is used

instances of the word "might"

No Comments

No Known Activations