INDEX

Explanations

references to comic book issue numbers in a specific format

oai_token-act-pair · gpt-3.5-turbo

references to comic book issues and serial numbers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

omorphic

-0.75

 fitness

-0.72

 cardio

-0.69

 basil

-0.68

 Sylv

-0.68

 laun

-0.68

 circulation

-0.67

 minded

-0.66

 contrace

-0.65

 provoking

-0.65

POSITIVE LOGITS

DIV

1.26

################################

1.20

###

1.14

########

1.13

################

0.94

!/

0.94

NB

0.90

RH

0.88

TOP

0.88

âĢİ

0.86

Activations Density 0.017%

references to comic book issue numbers in a specific format

references to comic book issues and serial numbers

No Comments

No Known Activations