INDEX

Explanations

titles and ranks related to authority or academic achievements

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

-0.69

 Biôgrafia

-0.68

MessageTagHelper

-0.59

 propOrder

-0.58

 nahilalakip

-0.57

InputTagHelper

-0.57

 ſch

-0.56

 kaarangay

-0.55

Manbalar

-0.54

 zeitung

-0.54

POSITIVE LOGITS

0.40

 strconv

0.39

Dr

0.38

льевич

0.35

0.34

Kol

0.34

</thead>

0.34

 SDLK

0.34

 rowspan

0.33

Activations Density 0.005%

titles and ranks related to authority or academic achievements

No Comments

No Known Activations