INDEX

Explanations

technical terms and phrases related to quantitative analysis and experimental procedures

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

signal

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

 kasarigan

-0.59

runApp

-0.42

:+:

-0.40

 autorytatywna

-0.40

 hObject

-0.39

Parcelize

-0.38

__/

-0.37

__))

-0.36

]{\

-0.36

StreetMap

-0.36

POSITIVE LOGITS

rungsseite

0.50

InSection

0.42

tvguidetime

0.40

 oprot

0.40

 bunu

0.40

ⓧ

0.40

 admitir

0.39

iseite

0.39

 للمعارف

0.38

SuspendLayout

0.38

Activations Density 0.166%

technical terms and phrases related to quantitative analysis and experimental procedures

signal

No Comments

No Known Activations