INDEX

Explanations

references to data organization and processing techniques

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

new or existing conditions

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

Filmographie

-0.38

 Interest

-0.38

 المعيارى

-0.36

Gob

-0.35

Inf

-0.35

Nar

-0.35

Interest

-0.35

fila

-0.34

awtextra

-0.34

SuccessListener

-0.34

POSITIVE LOGITS

 lenker

0.74

 propOrder

0.56

 unknownFields

0.55

IntoConstraints

0.49

 navideño

0.47

 régulière

0.46

 muerta

0.45

 navideña

0.45

 aikana

0.45

kijken

0.44

Activations Density 0.166%

references to data organization and processing techniques

new or existing conditions

No Comments

No Known Activations