INDEX

Explanations

terms related to performance and capability assessments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.36.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

omitempty

-0.42

 createSlice

-0.41

plona

-0.40

PickerController

-0.39

ContextCompat

-0.39

 actionMode

-0.38

DebuggerStep

-0.38

どうも

-0.38

Plus

-0.36

bindingNavigator

-0.36

POSITIVE LOGITS

 equally

2.59

 Equally

2.11

 altrett

1.91

 genauso

1.80

 igualmente

1.69

 ebenso

1.59

 stejně

1.59

 igual

1.52

equ

1.46

 equal

1.43

Activations Density 0.680%

terms related to performance and capability assessments

No Comments

No Known Activations