INDEX

Explanations

HTML attribute assignments

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

references to figures and tables

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

ActionCreators

-0.69

ագրություններ

-0.63

protoimpl

-0.60

 يتيمه

-0.58

 ModelExpression

-0.56

 Biôgrafia

-0.54

 ویکی‌پدیا

-0.54

 onCreateView

-0.54

fjspx

-0.54

usermodel

-0.53

POSITIVE LOGITS

glTex

0.43

0.42

 right

0.41

 certified

0.40

mäßigen

0.38

 droits

0.35

 licensed

0.35

 Roll

0.35

 access

0.33

 health

0.33

Activations Density 0.009%

HTML attribute assignments

references to figures and tables

No Comments

No Known Activations