INDEX

Explanations

adjectives related to minimalism or essential elements

oai_token-act-pair · gpt-3.5-turbo

references to the concept of being minimal or stripped down

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

anwhile

-0.85

abulary

-0.76

arios

-0.74

 Democr

-0.72

oldown

-0.71

afort

-0.70

ickr

-0.70

agents

-0.68

å°Ĩ

-0.68

alez

-0.67

POSITIVE LOGITS

footed

1.17

foot

1.16

bones

1.08

faced

1.03

feet

0.99

clad

0.96

handed

0.94

 bare

0.94

teen

0.93

chest

0.91

Activations Density 0.014%

adjectives related to minimalism or essential elements

references to the concept of being minimal or stripped down

No Comments

No Known Activations