INDEX

Explanations

references to exercise routines and workouts

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 disambiguazione

-0.68

 CURIAM

-0.66

Parcelize

-0.59

ellido

-0.59

 famílias

-0.59

 lápis

-0.57

 mães

-0.57

argli

-0.54

 digitais

-0.54

 zelve

-0.54

POSITIVE LOGITS

 workout

1.05

workout

1.00

Workout

0.97

 Workout

0.96

 workouts

0.87

training

0.60

 raiſ

0.57

 exerci

0.56

fitness

0.54

 exerc

0.54

Activations Density 0.002%

references to exercise routines and workouts

No Comments

No Known Activations