INDEX

Explanations

instances of scientific terminology related to research studies and methodologies

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

Tokens in identifiers, such as URLs or filenames

np_acts-logits-general · gemini-2.0-flash

pone and citation numbers

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

niſſe

-1.33

parsedMessage

-1.33

ſicht

-1.32

<unused8>

-1.30

<unused14>

-1.30

[@BOS@]

-1.30

<unused74>

-1.30

<unused79>

-1.30

<unused41>

-1.30

<unused16>

-1.30

POSITIVE LOGITS

↵↵

0.41

 dieux

0.32

 lagt

0.27

 esprits

0.27

↵

0.27

 Königin

0.26

 enfans

0.26

 gibi

0.25

得

0.25

 Wissens

0.25

Activations Density 1.648%

instances of scientific terminology related to research studies and methodologies

Tokens in identifiers, such as URLs or filenames

pone and citation numbers

No Comments

No Known Activations