INDEX

Explanations

structured elements and function definitions in programming or analytical contexts

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

correctness and facts

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

sphase

-0.38

FormTagHelper

-0.34

ModelSerializer

-0.33

ikations

-0.31

 videojuegos

-0.31

 dasarnya

-0.31

 phase

-0.30

 dasar

-0.30

{#

-0.29

 Extended

-0.29

POSITIVE LOGITS

 فريبيس

0.68

 استنادى

0.55

GEBURTS

0.54

:✨

0.54

 يتيمه

0.52

 TestBed

0.52

Tikang

0.51

qrstuvwxyz

0.50

ylvan

0.47

 Савезне

0.47

Activations Density 0.357%

structured elements and function definitions in programming or analytical contexts

correctness and facts

No Comments

No Known Activations