INDEX

Explanations

rooms or spaces that are full of something

oai_token-act-pair · gpt-3.5-turbo

references to rooms or spaces filled with items or people

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

 Disciple

-0.79

pher

-0.71

 Pere

-0.71

apters

-0.67

Downloadha

-0.66

ucl

-0.66

ovan

-0.66

oice

-0.66

gemony

-0.64

 Calder

-0.64

POSITIVE LOGITS

screen

0.98

erton

0.83

heartedly

0.81

ening

0.79

 blown

0.79

grown

0.77

width

0.77

 grown

0.76

ness

0.75

 moon

0.75

Activations Density 0.028%

rooms or spaces that are full of something

references to rooms or spaces filled with items or people

No Comments

No Known Activations