INDEX

Explanations

references to code structure and Java components

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

java class constructor

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

SequentialGroup

-0.68

 dAtA

-0.64

fjspx

-0.62

 ब्रेकडाउन

-0.61

ParallelGroup

-0.57

yyb

-0.55

⭗

-0.54

[@BOS@]

-0.53

<unused52>

-0.53

AnchorStyles

-0.53

POSITIVE LOGITS

RefNanny

0.37

Einzelnachweise

0.35

 mères

0.34

 épaules

0.34

 judicia

0.33

Judicial

0.33

 Aktu

0.33

 boîtes

0.32

 boî

0.31

 plonge

0.31

Activations Density 0.001%

references to code structure and Java components

java class constructor

No Comments

No Known Activations