INDEX

Explanations

words related to the physical condition of being worn out or damaged

oai_token-act-pair · gpt-3.5-turbo

references to wearing and the condition of worn items

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

aminer

-0.93

raction

-0.68

izoph

-0.68

yrinth

-0.66

akespeare

-0.66

à¼

-0.64

ribune

-0.64

ulhu

-0.62

llor

-0.62

ommod

-0.62

POSITIVE LOGITS

 worn

1.06

ables

0.92

wear

0.85

 apparel

0.84

IGH

0.84

nesday

0.82

bands

0.81

bones

0.81

 Wear

0.79

pins

0.79

Activations Density 0.021%

words related to the physical condition of being worn out or damaged

references to wearing and the condition of worn items

No Comments

No Known Activations