INDEX

Explanations

specific measurements and classifications related to property or spaces

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 EconPapers

-0.48

شهاد

-0.48

 nhiên

-0.39

RectangleBorder

-0.37

entary

-0.36

 thác

-0.36

듯

-0.35

tiary

-0.35

Rebs

-0.34

 Goal

-0.34

POSITIVE LOGITS

 Brooklyn

0.55

NYC

0.54

Brooklyn

0.53

NYC

0.51

 autorytatywna

0.51

 Bronx

0.47

AnimationsModule

0.47

Manhattan

0.47

 للاسماء

0.46

UnsafeEnabled

0.46

Activations Density 0.000%

specific measurements and classifications related to property or spaces

No Comments

No Known Activations