INDEX

Explanations

terms related to research studies and their results

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

ranked damage, treatment group

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

ⓧ

-0.34

 propOrder

-0.31

 gyhoeddwyd

-0.30

 gainera

-0.30

клопе

-0.29

Стан

-0.29

다

-0.29

Obrigada

-0.29

förs

-0.29

 tilgjenge

-0.29

POSITIVE LOGITS

 ProtoMessage

0.93

 NSCoder

0.64

 intStringLen

0.63

invokeLater

0.60

enderror

0.59

WithMany

0.59

moveToFirst

0.59

رشف

0.58

detectChanges

0.54

ſicht

0.54

Activations Density 0.357%

terms related to research studies and their results

ranked damage, treatment group

No Comments

No Known Activations