INDEX

Explanations

phrases related to potential actions and their implications

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

Follows auxiliary verbs (will, can, would, may)

np_acts-logits-general · gemini-2.0-flash

modal verb followed by action verb

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.22.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

 ruining

-0.68

 intervened

-0.60

 chofe

-0.57

 provocando

-0.57

 inflicting

-0.56

 enfans

-0.56

 begle

-0.56

 endangering

-0.56

 scaring

-0.56

 spoiling

-0.56

POSITIVE LOGITS

 result

0.64

 further

0.63

 greatly

0.59

})`

0.58

VersionUID

0.57

)$\\

0.55

unknownFields

0.55

 ultimately

0.55

出版年

0.55

 results

0.55

Activations Density 0.471%

phrases related to potential actions and their implications

Follows auxiliary verbs (will, can, would, may)

modal verb followed by action verb

No Comments

No Known Activations