INDEX

Explanations

programming-related keywords and object method calls

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

Code snippets with arrows and colons

np_acts-logits-general · gemini-2.0-flash

object property access

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

AndroidJUnit

-0.62

 ujednoznacz

-0.59

GenerationType

-0.56

RTLI

-0.56

kwds

-0.55

 betweenstory

-0.54

 يتيمه

-0.52

UnusedPrivate

-0.51

cæ

-0.49

oa̍t

-0.49

POSITIVE LOGITS

 createState

0.45

 internal

0.40

elesai

0.39

 total

0.38

 formalin

0.37

 película

0.36

 Kitt

0.36

 interpolated

0.36

estroyer

0.36

 direct

0.36

Activations Density 0.250%

programming-related keywords and object method calls

Code snippets with arrows and colons

object property access

No Comments

No Known Activations