INDEX

Explanations

references to academic studies and their findings

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

discussing research findings

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

ItemBackground

-0.42

tableFuture

-0.41

 Taktlose

-0.39

 Audiodateien

-0.38

 initComponents

-0.37

 kink

-0.37

webElementXpaths

-0.35

ChildScrollView

-0.34

SharedCtor

-0.34

 stretch

-0.34

POSITIVE LOGITS

UnusedPrivate

0.50

 kasarigan

0.44

violi

0.43

ต่างๆ

0.43

OCCURRED

0.42

PYX

0.42

 BaseActivity

0.42

どれも

0.41

sertations

0.41

Доброго

0.41

Activations Density 0.080%

references to academic studies and their findings

discussing research findings

No Comments

No Known Activations