INDEX

Explanations

instances of numbers or words related to measurements or quantities

oai_token-act-pair · gpt-3.5-turbo

textual elements and phrases indicating conflict or contrast

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Carey

-0.89

 Shan

-0.84

HTC

-0.84

uren

-0.83

Mur

-0.83

 Miche

-0.81

ĨĴ

-0.81

 Micha

-0.80

 savage

-0.78

 Brist

-0.77

POSITIVE LOGITS

 Ball

2.49

Ball

2.41

 ball

2.23

 BALL

2.17

ball

2.06

 Balls

2.02

 balls

2.01

BALL

1.90

balls

1.71

 Balloon

1.32

Activations Density 0.193%

instances of numbers or words related to measurements or quantities

textual elements and phrases indicating conflict or contrast

No Comments

No Known Activations