INDEX

Explanations

references to presentations and discussions about research or results

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

 AspNetCore

-0.48

 Meksiku

-0.47

 dezelve

-0.43

RotationOrder

-0.42

camps

-0.41

 erſt

-0.41

 noDo

-0.40

 hvě

-0.40

 ſind

-0.40

 AdapterView

-0.39

POSITIVE LOGITS

 slides

0.57

 presentation

0.50

slides

0.48

 presentations

0.46

twimg

0.45

 Presentation

0.44

 Slides

0.41

Slide

0.41

Overview

0.41

 Presentations

0.41

Activations Density 0.013%

references to presentations and discussions about research or results

No Comments

No Known Activations