INDEX

Explanations

references back to specific studies or data points within a set of research or analysis

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

reponse fausse

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

 staff

-0.32

 Tapia

-0.32

 aimer

-0.30

Staff

-0.29

 Escobar

-0.28

 Staff

-0.28

лец

-0.28

 consectetur

-0.28

järvi

-0.28

 Espinoza

-0.27

POSITIVE LOGITS

 CreateTagHelper

0.69

 للمعارف

0.66

rungsseite

0.64

:✨

0.62

 Signalez

0.62

 surla

0.61

évaluateur

0.59

 typelib

0.59

ftagPool

0.59

 Meksiku

0.56

Activations Density 0.192%

references back to specific studies or data points within a set of research or analysis

reponse fausse

No Comments

No Known Activations