INDEX

Explanations

structures related to graph representations

oai_token-act-pair · gpt-4o-mini Triggered by @bot

foreign and technical terms

np_acts-logits-general · gemini-2.5-flash-lite

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

WriteTagHelper

-0.50

ſelf

-0.48

ſelves

-0.48

 насељу

-0.47

stdc

-0.46

IGENCE

-0.44

mindedness

-0.43

expandindo

-0.43

 AppCompatTheme

-0.43

tagHelperRunner

-0.43

POSITIVE LOGITS

hdashline

0.48

Cyfarwyddwr

0.45

 EconPapers

0.43

MathML

0.43

 mijne

0.42

 initComponents

0.41

 tyle

0.40

ölker

0.39

 imaginación

0.38

 dagens

0.37

Activations Density 0.027%

structures related to graph representations

foreign and technical terms

No Comments

No Known Activations