INDEX

Explanations

key pieces of information related to health and safety assessments

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

cause undetermined

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

:✨

-0.57

vellous

-0.56

 autorytatywna

-0.51

 BIBSYS

-0.49

 gröss

-0.47

 randomised

-0.47

 Neigh

-0.47

Neighbours

-0.46

 apologised

-0.45

Personensuche

-0.43

POSITIVE LOGITS

doPost

0.44

 StatefulWidget

0.43

 chemical

0.40

TSCA

0.38

AutoField

0.38

最快更新

0.38

inture

0.37

eval

0.37

seins

0.36

 emergency

0.36

Activations Density 0.001%

key pieces of information related to health and safety assessments

cause undetermined

No Comments

No Known Activations