INDEX

Explanations

references to scientific concepts, methodologies, and variables in academic writing

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

these noun

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

 esternos

-0.74

utilisons

-0.59

SBATCH

-0.57

 صوتيه

-0.56

 brevis

-0.55

AISSEE

-0.54

thyst

-0.48

 calendriers

-0.48

chism

-0.48

 صفحۀ

-0.48

POSITIVE LOGITS

 AssemblyCulture

0.63

principalTable

0.42

 ErrIntOverflow

0.40

jsonwebtoken

0.39

KommentareTeilen

0.36

staande

0.34

DoubleQuotes

0.33

demás

0.31

 drinks

0.31

ArgsConstructor

0.31

Activations Density 0.250%

references to scientific concepts, methodologies, and variables in academic writing

these noun

No Comments

No Known Activations