INDEX

Explanations

the concept of counting or enumerating more of something

oai_token-act-pair · gpt-3.5-turbo

instances of the word "more."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

efe

-0.77

both

-0.77

iman

-0.72

abus

-0.72

anship

-0.71

obar

-0.71

odor

-0.70

Gra

-0.70

ister

-0.67

alogy

-0.67

POSITIVE LOGITS

 than

1.06

 rounds

0.93

 installments

0.90

 years

0.86

 weeks

0.86

 months

0.84

 inches

0.82

 digits

0.82

 kilometres

0.79

 iterations

0.78

Activations Density 0.051%

the concept of counting or enumerating more of something

instances of the word "more."

No Comments

No Known Activations