INDEX

Explanations

specific numerical or coded references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 5-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_5/width_16k/average_l0_68

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.5.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

andescent

-0.51

 portée

-0.45

ιά

-0.45

JHEP

-0.44

Халык

-0.43

스크

-0.42

olkata

-0.42

 estrema

-0.42

 Pickering

-0.41

 känd

-0.41

POSITIVE LOGITS

Aiheesta

0.82

OGND

0.78

endphp

0.73

 مشين

0.70

 Италијани

0.69

exitRule

0.68

 himo

0.66

 ويكيپيديا

0.64

 AssemblyTitle

0.63

Бахар

0.62

Activations Density 1.015%

specific numerical or coded references

No Comments

No Known Activations