INDEX

Explanations

numerical data or figures related to scientific research

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

citations and numbers

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

 auroit

-0.41

 pouvoit

-0.41

 extérieures

-0.41

LOYEES

-0.41

gamesh

-0.40

IndentedString

-0.39

lehre

-0.39

 vägen

-0.37

 extérieurs

-0.37

rasında

-0.37

POSITIVE LOGITS

 betweenstory

0.64

RegressionTest

0.64

ValueStyle

0.61

orianCalendar

0.53

 <<<<<<<<<<<<<<

0.52

Ket

0.47

ⓧ

0.46

期刊论文

0.46

transQ

0.45

Question

0.44

Activations Density 0.023%

numerical data or figures related to scientific research

citations and numbers

No Comments

No Known Activations