INDEX

Explanations

the number zero being mentioned

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "zero."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

Downloadha

-0.91

soDeliveryDate

-0.77

rough

-0.75

icken

-0.74

areth

-0.72

ournals

-0.72

asio

-0.70

lov

-0.70

Quote

-0.70

ournal

-0.68

POSITIVE LOGITS

 zero

0.87

zero

0.81

 Hedge

0.81

 Escape

0.76

 amount

0.73

sum

0.70

 tolerance

0.69

ebra

0.67

xd

0.67

 percent

0.67

Activations Density 0.018%

the number zero being mentioned

occurrences of the word "zero."

No Comments

No Known Activations