INDEX

Explanations

phrases that qualify or clarify the accuracy and representation of information

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

actual real final theoretical

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

 tradição

-0.34

 agência

-0.32

 Zone

-0.31

lò

-0.31

blog

-0.30

prz

-0.30

ocarcinoma

-0.29

 comprome

-0.29

 kiss

-0.29

ophores

-0.29

POSITIVE LOGITS

 nakalista

0.70

 reality

0.64

 actuality

0.63

httphttps

0.62

 EconPapers

0.60

tanleria

0.60

 réalité

0.59

 فريبيس

0.59

Manbalar

0.58

المكان

0.57

Activations Density 0.068%

phrases that qualify or clarify the accuracy and representation of information

actual real final theoretical

No Comments

No Known Activations