INDEX

Explanations

references to names and parties involved in legal cases

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

names followed by a period

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

 tough

-0.30

呢

-0.29

再

-0.29

 triliun

-0.29

◀

-0.28

 fine

-0.28

ngdoc

-0.27

big

-0.27

[*]

-0.27

 guste

-0.27

POSITIVE LOGITS

rungsseite

0.84

parsedMessage

0.76

 ſch

0.68

ſelf

0.67

 ſol

0.66

ſel

0.66

 ſta

0.65

󰀄

0.64

 myſelf

0.64

$_"

0.64

Activations Density 0.229%

references to names and parties involved in legal cases

names followed by a period

No Comments

No Known Activations