INDEX

Explanations

statistics and percentages related to demographics and health

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

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

 risk

-0.31

 chú

-0.31

 slightly

-0.28

 further

-0.28

วรร

-0.27

 Fifth

-0.27

 bold

-0.26

だったが

-0.26

bit

-0.25

 direct

-0.25

POSITIVE LOGITS

rungsseite

0.80

новниш

0.76

exitRule

0.76

Дереккөздер

0.75

UserScript

0.74

 kasarigan

0.72

 Signalez

0.70

RegressionTest

0.68



0.66

 validamos

0.66

Activations Density 0.137%

statistics and percentages related to demographics and health

No Comments

No Known Activations