INDEX

Explanations

phrases related to legal terminology and procedures

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

phrases preceded by "in part" or "words"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

 wireType

-0.52

EndInit

-0.45

íncia

-0.41

localctx

-0.41

󠁮

-0.40

Museo

-0.39

 Infórmanos

-0.38

POM

-0.38

exitRule

-0.37

haikusbot

-0.37

POSITIVE LOGITS

setVerticalGroup

0.48

0.45

 “…

0.45

toThrow

0.44

“

0.43

harapkan

0.42

 Describes

0.42

 descriptions

0.41

 mówią

0.41

berdayakan

0.41

Activations Density 0.170%

phrases related to legal terminology and procedures

phrases preceded by "in part" or "words"

No Comments

No Known Activations