INDEX

Explanations

proper nouns, particularly names of people

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Einer

-0.42

 Hause

-0.37

发表于

-0.36

hdashline

-0.36

:]:

-0.35

really

-0.34

 autorytatywna

-0.34

XmlEnum

-0.33

 really

-0.33

сос

-0.32

POSITIVE LOGITS

 فريبيس

0.72

 dezelve

0.66

ähteet

0.65

 يتيمه

0.63

iſen

0.63

Gön

0.60

AddTagHelper

0.59

VYMaps

0.58

épend

0.58

CppCodeGen

0.56

Activations Density 0.042%

proper nouns, particularly names of people

No Comments

No Known Activations