INDEX

Explanations

mathematical notation and symbols, particularly those related to functions and sequences

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 <=",

-0.53

 surla

-0.51

WireFormat

-0.49

UnusedPrivate

-0.48

EDEFAULT

-0.46

ofire

-0.43

protoimpl

-0.43

munk

-0.42

 Männern

-0.42

LOAT

-0.42

POSITIVE LOGITS

 previous

0.39

 AppComponent

0.38

鰭

0.36

 Previous

0.35

这时

0.34

 earlier

0.34

 eerder

0.34

 wcześ

0.33

 sebelumnya

0.33

يميديا

0.32

Activations Density 0.034%

mathematical notation and symbols, particularly those related to functions and sequences

No Comments

No Known Activations