INDEX

Explanations

references to academic or research concepts and terminology

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

citation details

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

сылкі

-0.38

 either

-0.27

thme

-0.26

mtrl

-0.26

FunctionFlags

-0.25

 חיצוניים

-0.25

tonsoft

-0.25

Сылтамалар

-0.25

 externi

-0.25

jdt

-0.24

POSITIVE LOGITS

0.62

+#+

0.57

Deskripsi

0.55



0.55

 Growing

0.52

Growing

0.51

HomeAsUp

0.51

ActionCreators

0.50

Specifications

0.50

WriteLiteral

0.50

Activations Density 0.263%

references to academic or research concepts and terminology

citation details

No Comments

No Known Activations

references to academic or research concepts and terminology

citation details

No Comments

No Known Activations