INDEX

Explanations

specific scientific terminology related to research findings or data

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

news reports and statements

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

AddTagHelper

-0.52

utilisons

-0.49

 CreateTagHelper

-0.44

 nahilalakip

-0.44

balleur

-0.43

phazard

-0.43

 Comprometido

-0.42

ActionCreators

-0.41

 Exactos

-0.40

Tikang

-0.40

POSITIVE LOGITS

 comment

0.67

 comments

0.59

 statement

0.56

 reports

0.55

 report

0.54

 informes

0.53

 հղումներ

0.52

abestanden

0.52

 komment

0.50

 accusation

0.49

Activations Density 0.186%

specific scientific terminology related to research findings or data

news reports and statements

No Comments

No Known Activations