INDEX

Explanations

elements and attributes from structured data and code snippets

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

codes and numbers

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

 initComponents

-0.59

 disambiguazione

-0.57

Хьажоргаш

-0.55

 Италијани

-0.52

ViewFeatures

-0.52

featureID

-0.51

transQ

-0.49

awtextra

-0.49

 linkovi

-0.49

iprot

-0.48

POSITIVE LOGITS

 unhealthy

0.47

RemoveField

0.44

icy

0.43

 tiny

0.43

 worrying

0.42

 bread

0.42

 Hann

0.42

 phony

0.41

 violate

0.41

น้อย

0.41

Activations Density 0.011%

elements and attributes from structured data and code snippets

codes and numbers

No Comments

No Known Activations