INDEX

Explanations

words related to software components and licensing terms

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

lass

-0.35

 obsidian

-0.34

 Lasse

-0.32

 Obsidian

-0.31

 vérité

-0.31

illigen

-0.31

cli

-0.31

address

-0.30

ophon

-0.30

 salute

-0.30

POSITIVE LOGITS

Personendaten

0.82

 OMITBAD

0.77

oa̍t

0.74

 ویکی‌پدی

0.74

 AssemblyTitle

0.74

хьтан

0.73

JsonHelper

0.69

sizeCache

0.66

resave

0.65

WebVitals

0.64

Activations Density 0.011%

words related to software components and licensing terms

No Comments

No Known Activations