INDEX

Explanations

specific suffixes and prefixes in words, particularly those suggesting a relationship or direction

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

from japanese english korean

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

Not in Any Lists

No Comments

Negative Logits

 informée

-0.46

 AppCompat

-0.40

 autorytatywna

-0.39

bonucleic

-0.39

HasColumnType

-0.38

OGND

-0.38

tettu

-0.37

itespace

-0.36

 navideñas

-0.36

Бахар

-0.35

POSITIVE LOGITS



0.49

hez

0.46

 raiſ

0.45

sern

0.45

에서

0.43

 houſe

0.42

 Foreign

0.42

ⓧ

0.42

 에서

0.42

ErrUnexpectedEOF

0.41

Activations Density 0.039%

specific suffixes and prefixes in words, particularly those suggesting a relationship or direction

from japanese english korean

No Comments

No Known Activations