INDEX

Explanations

comparative statements and analyses involving potential, effectiveness, and methodologies

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

comparing things

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

writeValue

-0.34

nete

-0.34

eren

-0.34

 رف

-0.33

’

-0.31

ng

-0.31

bnf

-0.31

亘

-0.31

ang

-0.30

套

-0.30

POSITIVE LOGITS

 disambiguazione

0.71

 ujednoznacz

0.69

SharedCtor

0.68

principalTable

0.65

fjspx

0.63

 transfieras

0.63

IsContent

0.62

 HttpNotFound

0.61

 Administrativna

0.60

 للمعارف

0.60

Activations Density 0.073%

comparative statements and analyses involving potential, effectiveness, and methodologies

comparing things

No Comments

No Known Activations