INDEX

Explanations

items related to coding or data structures

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

dictionary keys

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

 trekken

-0.48

 pouvoit

-0.43

 "','

-0.41

)))),

-0.40

 becauſe

-0.39

labus

-0.39

tadiene

-0.37

viewDidLoad

-0.36

 asistencia

-0.36

paravant

-0.36

POSITIVE LOGITS

 Chwiliwch

0.65

0.57

 Paglinawan

0.54

Бахар

0.52

nameof

0.51

хьтан

0.51

 gyhoeddwyd

0.49

RTEE

0.47

ⓧ

0.47

Datuak

0.46

Activations Density 0.010%

items related to coding or data structures

dictionary keys

No Comments

No Known Activations