INDEX

Explanations

patterns related to the word "yo" followed by a number representing some level or quantity

oai_token-act-pair · gpt-3.5-turbo

repeated instances of the term "yo."

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

ioned

-0.77

icates

-0.71

pillar

-0.71

pring

-0.65

icating

-0.64

IUM

-0.63

sound

-0.62

itative

-0.62

poke

-0.61

sworth

-0.61

POSITIVE LOGITS

vernment

1.03

achu

0.88

wu

0.77

yo

0.76

olicy

0.75

UGE

0.72

ignt

0.70

aeper

0.69

oths

0.67

yond

0.66

Activations Density 0.015%

patterns related to the word "yo" followed by a number representing some level or quantity

repeated instances of the term "yo."

No Comments

No Known Activations