INDEX

Explanations

conjunctions, transition words, and phrases indicating cause and effect in the 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

 Италијани

-0.93

 Roskov

-0.70

SourceChecksum

-0.70

ьаж

-0.69

 queſta

-0.68

 kaarangay

-0.68

 صوتيه

-0.67

 kasarigan

-0.66

İstinadlar

-0.66

出版年

-0.66

POSITIVE LOGITS

initComponents

0.39

また

0.37

principalTable

0.37

 juga

0.37

 ayrıca

0.35

Bagi

0.35

 veran

0.34

 también

0.33

Apakah

0.33

Activations Density 0.027%

conjunctions, transition words, and phrases indicating cause and effect in the text

No Comments

No Known Activations