INDEX

Explanations

phrases related to breaking into something (e.g., a home, computer, system)

oai_token-act-pair · gpt-3.5-turbo

the phrase "break into."

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

ennis

-0.72

entimes

-0.72

erity

-0.65

 qualified

-0.65

 timetable

-0.64

 postponed

-0.64

rious

-0.62

rolley

-0.61

rising

-0.60

bl

-0.59

POSITIVE LOGITS

ochet

0.77

mson

0.76

Sax

0.71

 adulthood

0.70

 burgl

0.68

Wall

0.67

taining

0.65

hibited

0.64

ework

0.63

ode

0.62

Activations Density 0.032%

phrases related to breaking into something (e.g., a home, computer, system)

the phrase "break into."

No Comments

No Known Activations