INDEX

Explanations

Extraneous whitespace and formatting artifacts in text

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

ages and years

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

ConstraintMaker

-0.52

GEBURTSDATUM

-0.52

 Roskov

-0.48

WriteLiteral

-0.43

 trun

-0.43

 tiguan

-0.41

pin

-0.41

ACHUSET

-0.40

Noc

-0.40

OGND

-0.39

POSITIVE LOGITS

 Inscrivez

0.52

 setempat

0.52

 stră

0.49

 titular

0.47

 Suivez

0.46

freopen

0.46

lapsingToolbar

0.45

ContextCompat

0.45

umumkan

0.44

 Bupati

0.42

Activations Density 0.018%

Extraneous whitespace and formatting artifacts in text

ages and years

No Comments

No Known Activations