INDEX

Explanations

instances of numerical expressions and related contextual words

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

articles and numbers followed by specific subjects

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

 darse

-0.45

ArrowToggle

-0.41

SharedCtor

-0.41

 piú

-0.40

 darte

-0.36

 sarili

-0.35

SPARENT

-0.34

 miglia

-0.34

どころ

-0.34

 misure

-0.33

POSITIVE LOGITS

 مشين

0.56

رشف

0.49

enumii

0.48

发表于

0.47

UnusedPrivate

0.46

品

0.46

Noi

0.46

ipo

0.45

 ब्रेकडाउन

0.45

mpo

0.45

Activations Density 0.154%

instances of numerical expressions and related contextual words

articles and numbers followed by specific subjects

No Comments

No Known Activations