INDEX

Explanations

specific identifiers or names associated with people or roles

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

names and professions

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

хьтан

-0.59

 мәкал

-0.58

 BoxFit

-0.52

EnableWeb

-0.48

ainville

-0.48

PreExecute

-0.48

setDo

-0.46

posedge

-0.44

 OnTrigger

-0.44

OLEAN

-0.43

POSITIVE LOGITS

 Paglinawan

0.48

 profiles

0.48

MLLoader

0.44

 constitucional

0.44

同じく

0.43

 recrutement

0.43

 Савезне

0.42

profiles

0.40

 personalities

0.40

 biographies

0.39

Activations Density 0.029%

specific identifiers or names associated with people or roles

names and professions

No Comments

No Known Activations