INDEX

Explanations

code-related terms, particularly those associated with data representation and manipulation functions

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

string to hex

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

Personendaten

-0.82

parsedMessage

-0.81

+#+#

-0.77

endphp

-0.71

rungsseite

-0.71

IsMutable

-0.71

 snippetHide

-0.69

 שוליים

-0.68

әрмәләр

-0.66

 queſta

-0.66

POSITIVE LOGITS

ворот

0.36

hay

0.33

ロー

0.31

哭了

0.30

完

0.29

 onFailure

0.29

 dessous

0.29

specialchars

0.28

 Donnerstag

0.28

 terme

0.28

Activations Density 0.060%

code-related terms, particularly those associated with data representation and manipulation functions

string to hex

No Comments

No Known Activations