INDEX

Explanations

detailed descriptions and outlines of processes or methodologies

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

scientific or technical text involving mathematical or technical language, particularly words like "general", "terms", "method", and "approximation".

oai_token-act-pair · claude-3-5-haiku-20241022 Triggered by @neilrathi

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

 Entdecke

-0.39

 myſelf

-0.38

 ſta

-0.38

actually

-0.36

 ſtand

-0.36

tilhe

-0.34

 neceff

-0.34

 actually

-0.34

 rempliss

-0.33

 purpoſe

-0.33

POSITIVE LOGITS

 vague

0.62

 cursory

0.59

 vaguely

0.59

大致

0.58

Brief

0.57

 summaries

0.57

Rough

0.57

 rough

0.57

sơ

0.56

 propOrder

0.56

Activations Density 0.513%

detailed descriptions and outlines of processes or methodologies

scientific or technical text involving mathematical or technical language, particularly words like "general", "terms", "method", and "approximation".

No Comments

No Known Activations