INDEX

Explanations

patterns or sequences related to version numbers

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

Followed by a number or non-alphabetic character

np_acts-logits-general · gemini-2.0-flash

version numbers and punctuation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

GraphicsUnit

-0.82

-0.64

rd

-0.60

 oprot

-0.58

Kanpo

-0.57

omla

-0.57

تقاوى

-0.56

IsContent

-0.55

standers

-0.54

ddots

-0.53

POSITIVE LOGITS

ICEF

0.63

 donnée

0.56

(:,:,

0.54

gj

0.52

/');

0.51

(@"%@",

0.51

 souple

0.50

Referanser

0.50

何より

0.49

Konto

0.49

Activations Density 0.259%

patterns or sequences related to version numbers

Followed by a number or non-alphabetic character

version numbers and punctuation

No Comments

No Known Activations