INDEX

Explanations

code-related terms and methods for data parsing and serialization

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

 ujednoznacz

-0.49

 propOrder

-0.47

UnusedPrivate

-0.47

 CreateTagHelper

-0.46

 JpaRepository

-0.45

 TType

-0.42

 للاسماء

-0.41

setCancelable

-0.40

WritableDatabase

-0.40

\{\\

-0.39

POSITIVE LOGITS



0.48

JScrollPane

0.40

upol

0.39

瘩

0.38

 asupra

0.38

 parç

0.38

 AssemblyTitle

0.37

 pinulongan

0.37

 resolution

0.37

Portail

0.37

Activations Density 0.099%

code-related terms and methods for data parsing and serialization

No Comments

No Known Activations