INDEX

Explanations

descriptors related to effort, difficulty, or intensity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

 Monat

-0.49

 cejas

-0.46

 MenuItem

-0.45

 burbujas

-0.44

 Figure

-0.42

 forsø

-0.42

 bricolaje

-0.42

 disfrute

-0.41

ípio

-0.41

 Figuren

-0.40

POSITIVE LOGITS

Hard

0.90

 Hard

0.88

 faſt

0.86

 Hardin

0.78

soft

0.78

 HARD

0.77

Soft

0.77

 Fast

0.75

httphttps

0.75

 Soft

0.74

Activations Density 0.098%

descriptors related to effort, difficulty, or intensity

No Comments

No Known Activations