INDEX

Explanations

sequences that involve letters and characters in various combinations, potentially indicating patterns or codes

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

polynomial simplification and rearrangement

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

ſchaft

-0.95

 useAppContext

-0.92

ſehen

-0.88

 queſta

-0.86

 المعيارى

-0.86

niſſe

-0.85

 betweenstory

-0.85

httphttps

-0.84

iſchen

-0.83

OGND

-0.82

POSITIVE LOGITS

main

0.36

 Außerdem

0.33

 satunya

0.31

0.29

Außerdem

0.29

0.28

 inneces

0.27

 Sociales

0.27

ma

0.27

通

0.27

Activations Density 0.023%

sequences that involve letters and characters in various combinations, potentially indicating patterns or codes

polynomial simplification and rearrangement

No Comments

No Known Activations