INDEX

Explanations

references to news updates and information across various topics

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

news items

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

 ComVisible

-0.42

PSM

-0.41

rat

-0.41

REM

-0.39

ylde

-0.38

andExpect

-0.36

oluted

-0.36

Mate

-0.36

 mate

-0.35

formKey

-0.35

POSITIVE LOGITS

 news

1.01

 noticia

0.92

 noticias

0.90

 NEWS

0.86

 notícias

0.82

 tidings

0.81

 announcements

0.81

 notícia

0.77

news

0.76

 berita

0.75

Activations Density 0.016%

references to news updates and information across various topics

news items

No Comments

No Known Activations