INDEX

Explanations

instructions or recommendations to perform a specific action before proceeding with another action

oai_token-act-pair · gpt-3.5-turbo

instances of the word "before" indicating prior actions or considerations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãĤ·

-0.76

atic

-0.72

PsyNetMessage

-0.69

ãĤ¨

-0.69

artisan

-0.68

ãĥķãĤ©

-0.67

onet

-0.67

aic

-0.67

OTOS

-0.66

aez

-0.66

POSITIVE LOGITS

 proceeding

1.50

 committing

1.39

 deciding

1.27

 submitting

1.26

 entering

1.22

 embark

1.22

 attempting

1.19

 concluding

1.18

 heading

1.17

 applying

1.16

Activations Density 0.069%

instructions or recommendations to perform a specific action before proceeding with another action

instances of the word "before" indicating prior actions or considerations

No Comments

No Known Activations