INDEX

Explanations

references to the word "pie"

oai_token-act-pair · gpt-3.5-turbo

references to "pie."

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

iveness

-0.85

spring

-0.77

runners

-0.70

sers

-0.70

 Examiner

-0.69

iqueness

-0.68

nen

-0.67

 EDITION

-0.67

angers

-0.65

wip

-0.64

POSITIVE LOGITS

 pies

1.13

pie

1.11

ced

1.02

Pie

1.02

pie

1.01

cing

0.98

MpServer

0.95

 crust

0.95

Pie

0.93

 desserts

0.84

Activations Density 0.023%

references to the word "pie"

references to "pie."

No Comments

No Known Activations