INDEX

Explanations

names or identifiers related to individuals in academic or research contexts

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

initials and abbreviations starting with j

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

بوابة

-0.55

 skydd

-0.47

 BrowserModule

-0.47

abstractmethod

-0.46

AuthProvider

-0.45

 besky

-0.43

 rozwój

-0.42

 negó

-0.42

 Normdatei

-0.41

IBar

-0.41

POSITIVE LOGITS

 defaultstate

0.94

0.80

0.79

Jod

0.75

JI

0.74

JJ

0.73

Joc

0.72

Juf

0.71

 jLabel

0.70

Activations Density 0.229%

names or identifiers related to individuals in academic or research contexts

initials and abbreviations starting with j

No Comments

No Known Activations