INDEX

Explanations

coding-related terms and constructs, particularly those associated with object manipulation and data structures

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

InputBorder

-0.57

UnusedPrivate

-0.54

󠁴

-0.49

AndEndTag

-0.47

 BorderSide

-0.46

➟

-0.45

 surla

-0.43

AddHtmlAttribute

-0.43

delwed

-0.43

SBATCH

-0.42

POSITIVE LOGITS

 дописавши

0.40



0.40

RetentionPolicy

0.39

 Schrifts

0.38

Personendaten

0.38

 Lokales

0.35

újo

0.35

ContentAsync

0.34

 order

0.34

contentLoaded

0.33

Activations Density 0.023%

coding-related terms and constructs, particularly those associated with object manipulation and data structures

No Comments

No Known Activations