INDEX

Explanations

words and phrases related to scientific research, investigations, and conclusions

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

recent recent recenct

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

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

:✨

-0.69

Eventually

-0.67

 Eventually

-0.66

 bientôt

-0.65

 domani

-0.64

Tomorrow

-0.62

 eventually

-0.61

Soon

-0.61

eventually

-0.59

 tomorrow

-0.59

POSITIVE LOGITS

 since

1.80

 lately

1.65

since

1.58

 recently

1.54

 SINCE

1.51

 recent

1.46

recently

1.44

Since

1.33

 Recently

1.31

Recently

1.30

Activations Density 0.446%

words and phrases related to scientific research, investigations, and conclusions

recent recent recenct

No Comments

No Known Activations