INDEX

Explanations

terms reflecting emotional or subjective responses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 23-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_23/width_16k/average_l0_75

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.23.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

 purpoſe

-0.80

出版年

-0.79

 perſon

-0.77

 houſe

-0.76

 fubject

-0.75

 ſtate

-0.75

 perfons

-0.70

 fusca

-0.70

 pleaſure

-0.69

 OFDb

-0.69

POSITIVE LOGITS

 fuoco

0.49

—”

0.45

fjspx

0.44

 Chwiliwch

0.43

πάρχ

0.41

 Ersten

0.40

umph

0.40

ingly

0.39

 Escrit

0.37

 boletín

0.37

Activations Density 0.281%

terms reflecting emotional or subjective responses

No Comments

No Known Activations