INDEX

Explanations

references to data access and datasets in scientific contexts

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

data and downloads

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

pin

-0.30

 Margin

-0.29

man

-0.29

 isolated

-0.28

pin

-0.27

nbsp

-0.27

texttt

-0.27

IContainer

-0.27

 pins

-0.27

alk

-0.26

POSITIVE LOGITS

 <>",

0.93

 Administrativna

0.65

 Italij

0.65

 հղումներ

0.62

 müſſen

0.61

 समीक्षक

0.60

spodnie

0.60

 oprot

0.60

 indígen

0.59

majánló

0.59

Activations Density 0.079%

references to data access and datasets in scientific contexts

data and downloads

No Comments

No Known Activations