INDEX

Explanations

terms related to interactions among elements or variables

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 1-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_1/width_16k/average_l0_69

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.1.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

 Bubble

-0.47

 Roof

-0.44

 roof

-0.44

FlatAppearance

-0.41

 protoimpl

-0.41

思い

-0.40

뀔

-0.40

 bubble

-0.39

Roof

-0.38

Baz

-0.38

POSITIVE LOGITS

 interaction

0.81

Interaction

0.76

interaction

0.75

 Interaction

0.71

 ویکی‌پدیای

0.68

 interacción

0.65

interact

0.62



0.60

 interact

0.59

Walker

0.59

Activations Density 0.390%

terms related to interactions among elements or variables

No Comments

No Known Activations