INDEX

Explanations

elements and symbols associated with programming and code structure

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

code/formatting tokens

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

tvguidetime

-0.72

 ब्रेकडाउन

-0.71

 يتيمه

-0.69

featureID

-0.68

ftagPool

-0.65

Portail

-0.65

 ujednoznacz

-0.63

 autorytatywna

-0.61

 zuſammen

-0.60

 estekak

-0.60

POSITIVE LOGITS

Sooo

0.29

 expressions

0.29

("

0.27

 text

0.26

('

0.26

('

0.25

 commentary

0.25

 buzz

0.25

ie

0.25

 lash

0.24

Activations Density 0.050%

elements and symbols associated with programming and code structure

code/formatting tokens

No Comments

No Known Activations