INDEX

Explanations

programming constructs and outputs related to logging and debugging information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

DockStyle

-0.70

 MonoBehaviour

-0.68

ロウィン

-0.66

 pleaſure

-0.65

 queſta

-0.65

<unused68>

-0.64

<unused17>

-0.64

<unused52>

-0.64

[@BOS@]

-0.63

<unused14>

-0.63

POSITIVE LOGITS

Pil

0.30

que

0.30

ism

0.30

say

0.29

cout

0.29

 Nếu

0.29

productId

0.29

fromnode

0.28

Activations Density 0.117%

programming constructs and outputs related to logging and debugging information

No Comments

No Known Activations