INDEX

Explanations

technical terms and structures related to data handling and programming errors

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

]")]

-0.56



-0.50

httphttps

-0.50

 AssemblyCulture

-0.50

GOTREF

-0.48

 ModelExpression

-0.47

CodeAttribute

-0.46

 Menschheit

-0.45

 esercit

-0.45

rrggbb

-0.44

POSITIVE LOGITS

 sumpay

0.39

 ModelRenderer

0.38

Externé

0.37

 propOrder

0.36

воз

0.35

 Leck

0.35

 pret

0.35

$_['

0.35

 inoc

0.34

 best

0.34

Activations Density 0.339%

technical terms and structures related to data handling and programming errors

No Comments

No Known Activations