INDEX

Explanations

terms that indicate error or inaccuracies in data or measurements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 1-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_1/width_16k/average_l0_69

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.1.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

console

-0.36

してい

-0.35

ってい

-0.34

%;">

-0.34

mixin

-0.34

 allé

-0.33

 הול

-0.33

 argument

-0.33

 Syrie

-0.33

 féd

-0.33

POSITIVE LOGITS

 error

1.35

 Error

1.17

error

1.04

Error

0.94

 errors

0.84

 ERROR

0.81

 Errors

0.79

 errore

0.74

Errorf

0.73

 erro

0.73

Activations Density 0.201%

terms that indicate error or inaccuracies in data or measurements

No Comments

No Known Activations