INDEX

Explanations

numerical values and sequences or patterns in text

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

classification stages

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

 durabilidad

-0.45

queryInterface

-0.42

transQ

-0.41

🟤

-0.39

MENAFN

-0.38

disposing

-0.37

KURZBESCHREIBUNG

-0.37

ſelf

-0.36

 AspNetCore

-0.36

 onResponse

-0.35

POSITIVE LOGITS

 Fish

0.49

CloseOperation

0.48

Fish

0.46

 similarly

0.45

ViewFeatures

0.44

kloped

0.44

同样

0.43

同樣

0.43

Cold

0.43

PyTuple

0.43

Activations Density 0.092%

numerical values and sequences or patterns in text

classification stages

No Comments

No Known Activations