INDEX

Explanations

patterns related to testing and evaluation in programming or mathematical contexts

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

code identifiers

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

 بيها

-0.55

TagMode

-0.53

ChrTalk

-0.52



-0.51

Portail

-0.50

Bibliograf

-0.50

 calendriers

-0.49

arounds

-0.49

 Enclosure

-0.49

ferro

-0.48

POSITIVE LOGITS

SharedCtor

0.43

DoubleQuotes

0.38

 ostavi

0.37

 stesso

0.34

 AGAIN

0.33

 again

0.33

resave

0.32

holdet

0.32

'{@

0.32

again

0.32

Activations Density 0.324%

patterns related to testing and evaluation in programming or mathematical contexts

code identifiers

No Comments

No Known Activations