INDEX

Explanations

words related to questions and answers

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

*-ende, -enden, -gende words

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

ambién

-0.59

alakip

-0.58

 förr

-0.54

 OFDb

-0.52

 kän

-0.50

 Einladung

-0.50

 sánchez

-0.50

tagHelperRunner

-0.50

 épis

-0.49

 äls

-0.49

POSITIVE LOGITS

acting

0.63

ating

0.58

menden

0.56

ting

0.52

jící

0.52

cting

0.50

enden

0.50

ising

0.50

ing

0.48

ing

0.47

Activations Density 0.011%

words related to questions and answers

*-ende, -enden, -gende words

No Comments

No Known Activations