INDEX

Explanations

names of individuals, likely in a context related to their professional roles or activities

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

completing names and identifiers

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_24/width_131k/average_l0_97

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.24.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

↵↵

-0.48

-0.44

 Breen

-0.41

 вида

-0.41

-0.40

 Weller

-0.40

 Schiller

-0.39

 polu

-0.39

 Baer

-0.39

 Werth

-0.38

POSITIVE LOGITS

 فريبيس

0.87

 dezelve

0.78

 houſe

0.75

ientras

0.74

ſelf

0.73

 zijne

0.73

 ſta

0.73

iſen

0.73

 Anſ

0.72

 plufieurs

0.71

Activations Density 0.071%

names of individuals, likely in a context related to their professional roles or activities

completing names and identifiers

No Comments

No Known Activations