INDEX

Explanations

terms related to user input, satisfaction, and involvement in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

/**

-0.38

utnik

-0.30

orcid

-0.28

 ones

-0.27

 thing

-0.27

ֹת

-0.27

 Hälfte

-0.27

 género

-0.26

関係

-0.26

り返

-0.26

POSITIVE LOGITS

IntoConstraints

0.65

atility

0.60

UnsafeEnabled

0.60

lesque

0.58

 Perſ

0.58

 dieſe

0.56

inSlope

0.56

BagConstraints

0.56

 للمعارف

0.56

 Inſ

0.56

Activations Density 0.110%

terms related to user input, satisfaction, and involvement in various contexts

No Comments

No Known Activations