INDEX

Explanations

phrases related to cause and effect

oai_token-act-pair · gpt-3.5-turbo

phrases describing physical mechanisms or interactions

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

enty

-0.63

Across

-0.62

 Including

-0.60

Hundreds

-0.59

among

-0.58

 millionaires

-0.58

 famously

-0.58

dozen

-0.57

retty

-0.57

Numbers

-0.56

POSITIVE LOGITS

 your

0.92

SCP

0.91

 oneself

0.84

 plaintiff

0.75

you

0.73

 defendant

0.71

 sender

0.70

 wearer

0.70

 purchaser

0.70

 Plaintiff

0.70

Activations Density 0.538%

phrases related to cause and effect

phrases describing physical mechanisms or interactions

No Comments

No Known Activations