INDEX

Explanations

class names and related identifiers in programming languages

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

class definitions and methods

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

 kasarigan

-0.81

awtextra

-0.74

UnusedPrivate

-0.73

SharedDtor

-0.73

LEncoder

-0.73

AndEndTag

-0.71

 فريبيس

-0.70

RegressionTest

-0.69

 autorytatywna

-0.69

Personendaten

-0.67

POSITIVE LOGITS

’

0.33

 workshop

0.31

 originally

0.31

 adopted

0.31

 founded

0.30

 resident

0.30

 Entity

0.29

Север

0.29

 entity

0.28

 Resident

0.28

Activations Density 0.003%

class names and related identifiers in programming languages

class definitions and methods

No Comments

No Known Activations