INDEX

Explanations

terms related to bending or curving actions

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-2b-pt-res/layer_1/width_16k/average_l0_102

Prompts (Dashboard)

36,864 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

del

-0.51

Sal

-0.48

 bodem

-0.48

Drawing

-0.47

Джерела

-0.44

カウン

-0.44

 figyel

-0.43

leden

-0.43

loten

-0.43

 drew

-0.43

POSITIVE LOGITS

 bend

0.82

 repos

0.74

 Theſe

0.73

Bend

0.72

 تانيه

0.71

 Cæsar

0.71

 bends

0.70

 myſelf

0.69

 bent

0.68

 Huguen

0.68

Activations Density 2.319%

terms related to bending or curving actions

No Comments

No Known Activations