INDEX

Explanations

proper nouns and specific references to people, places, or titles

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

speaking or talking

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 31-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_31/width_131k/average_l0_109

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.31.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

 ModelExpression

-0.49

 Signalez

-0.44

PreferredItem

-0.43

 Paglinawan

-0.43

 تانيه

-0.42

хьтан

-0.41

 ब्रेकडाउन

-0.40

تقاوى

-0.39

省市镇

-0.38

AISSEE

-0.38

POSITIVE LOGITS

 speaking

2.66

speaking

2.23

 Speaking

2.13

Speaking

2.11

 talking

1.73

 hablando

1.65

 parlant

1.50

 speak

1.48

talking

1.48

 berbicara

1.45

Activations Density 0.040%

proper nouns and specific references to people, places, or titles

speaking or talking

No Comments

No Known Activations