INDEX

Explanations

words related to understanding and affirmation

oai_token-act-pair · gpt-3.5-turbo

words that indicate confirmation, affirmation, or agreement

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

fml

-0.68

chrom

-0.67

jri

-0.67

 fuss

-0.66

 challeng

-0.66

awa

-0.65

elf

-0.65

flix

-0.62

pid

-0.62

esm

-0.61

POSITIVE LOGITS

ments

1.89

ment

1.83

ations

1.63

ation

1.46

ement

1.42

ably

1.42

MENT

1.41

MENTS

1.36

ements

1.34

ATIONS

1.31

Activations Density 0.143%

words related to understanding and affirmation

words that indicate confirmation, affirmation, or agreement

No Comments

No Known Activations