INDEX

Explanations

architectural terms and design elements

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

organizational patterns

np_acts-logits-general · gemini-2.5-flash-lite

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

CloseOperation

-0.57

WithIOException

-0.52

ISupport

-0.51

IndentedString

-0.49

SBATCH

-0.48

tvguidetime

-0.47

 pleaſure

-0.46

 defire

-0.46

typeDefinition

-0.45

StoryboardSegue

-0.45

POSITIVE LOGITS

whole

0.39

 Chwiliwch

0.39

 répé

0.39

ύ

0.37

yntaxException

0.37

verifyException

0.35

 whole

0.35

 כול

0.34

]")]

0.34

AddTagHelper

0.33

Activations Density 0.108%

architectural terms and design elements

organizational patterns

No Comments

No Known Activations