INDEX

Explanations

numerical indicators such as percentages or quantities

oai_token-act-pair · gpt-3.5-turbo

phrases expressing scarcity or low quantity

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

agher

-0.66

 weaving

-0.63

obo

-0.62

NW

-0.62

Draft

-0.60

 Bust

-0.59

 widening

-0.59

dress

-0.58

irst

-0.58

alam

-0.58

POSITIVE LOGITS

 hundred

0.86

een

0.84

 dozen

0.80

 mortals

0.77

est

0.76

ones

0.74

eenth

0.73

orem

0.73

ishers

0.71

 thousand

0.71

Activations Density 0.026%

numerical indicators such as percentages or quantities

phrases expressing scarcity or low quantity

No Comments

No Known Activations