INDEX

Explanations

numerical values and their formatting in a text

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

vis

-0.41

 surface

-0.37

 Länder

-0.37

sw

-0.36

AMES

-0.36

 smoothness

-0.35

 viscous

-0.35

litos

-0.35

СКО

-0.35

 thought

-0.35

POSITIVE LOGITS

 town

0.64

Autoritní

0.60

 towns

0.55

 Rural

0.55

Rural

0.54

RectangleBorder

0.52

 miejsco

0.52

locally

0.52

:✨

0.51

 TOWN

0.51

Activations Density 0.002%

numerical values and their formatting in a text

No Comments

No Known Activations