INDEX

Explanations

measurements and quantities related to cooking recipes

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

cooking instructions

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

Personendaten

-0.52

nologies

-0.47

OGND

-0.43

getSeconds

-0.43

WithMany

-0.42

 للمعارف

-0.41

Aiheesta

-0.40

ittoria

-0.40

qtype

-0.39

/*

-0.38

POSITIVE LOGITS

 culinary

0.52

 recipes

0.51

 Cooking

0.49

 cooking

0.49

 instructions

0.47

canestro

0.47

 cookbook

0.47

 recipe

0.46

 cook

0.46

 kitchen

0.45

Activations Density 0.036%

measurements and quantities related to cooking recipes

cooking instructions

No Comments

No Known Activations