INDEX

Explanations

programming-related identifiers and function names

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

C code, memory allocation, strings

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

Diweddarwch

-0.57

 queſta

-0.57

transQ

-0.56

 EconPapers

-0.55

addContainerGap

-0.52

 pleaſure

-0.50

 stockbild

-0.49

buttonBar

-0.48

MemoryWarning

-0.48

 Infór

-0.47

POSITIVE LOGITS

########.

1.27

libft

1.22

:+:

0.59

 text

0.44

GIVEREF

0.43

 madly

0.40

Ft

0.39

putstr

0.39

GOTREF

0.39

 Durand

0.38

Activations Density 0.001%

programming-related identifiers and function names

C code, memory allocation, strings

No Comments

No Known Activations