INDEX

Explanations

references to training and related activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.4.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

Beb

-0.71

 KeyCode

-0.69

 Osbourne

-0.69

 Diocese

-0.68

Edd

-0.68

kadot

-0.65

NEC

-0.64

 heapq

-0.64

 Clarinet

-0.63

addFlags

-0.63

POSITIVE LOGITS

 training

2.47

 Training

2.16

Training

2.11

training

2.09

 TRAINING

1.94

TRAINING

1.86

 trainings

1.86

 Trainings

1.68

 trained

1.62

trained

1.56

Activations Density 0.055%

references to training and related activities

No Comments

No Known Activations