INDEX

Explanations

numerical references related to counts or quantities

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

numbers followed by nouns

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

LLocation

-0.58

worse

-0.52

 thiệu

-0.52

 ResponseEntity

-0.50

twice

-0.49

ukunft

-0.49

躇

-0.49

 queſto

-0.48

 hardly

-0.48

 worse

-0.48

POSITIVE LOGITS

 defaultstate

0.47

 jotka

0.46

 dört

0.42

 våra

0.42

 betweenstory

0.41

énario

0.39

 fleste

0.39

 wszyscy

0.39

 four

0.38

 übrigen

0.38

Activations Density 0.029%

numerical references related to counts or quantities

numbers followed by nouns

No Known Activations