INDEX

Explanations

repeated patterns or placeholders in 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

RenderAtEndOf

-0.92

 autorytatywna

-0.91

PyExc

-0.83

featureID

-0.80

RectangleBorder

-0.78

parsedMessage

-0.78

'))

-0.77

AnchorStyles

-0.77

+#+#

-0.74

EDEFAULT

-0.74

POSITIVE LOGITS

 coloré

0.49

 colorés

0.47

 cálido

0.45

 masiva

0.45

 creado

0.44

 nôtre

0.44

 spécialisé

0.43

 acogedor

0.43

 húmedo

0.43

 llenos

0.41

Activations Density 0.037%

repeated patterns or placeholders in text

No Comments

No Known Activations