INDEX

Explanations

symbols and punctuation indicating the structure of code or data

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

 Wikimedijinoj

-0.51

ReactDOM

-0.47

IonicModule

-0.45

ScopeManager

-0.45

CGRectMake

-0.44

 gynhyrchwyd

-0.42

PDOException

-0.41

Chham

-0.41

MLLoader

-0.40

mtrl

-0.39

POSITIVE LOGITS

fromnode

0.54

AndEndTag

0.43

 utafiti

0.40

 <=",

0.37

indd

0.37

bitat

0.36

 оригіналу

0.36

LabelTagHelper

0.36

Dad

0.36

 consentement

0.36

Activations Density 0.073%

symbols and punctuation indicating the structure of code or data

references to figures and tables

No Comments

No Known Activations