INDEX

Explanations

references to rankings or lists that indicate top positions or achievements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.4.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

nxt

-0.51

̀nh

-0.50

ERTS

-0.49

 Sonder

-0.48

field

-0.47

きた

-0.47

çük

-0.47

Zwischen

-0.46

 oltre

-0.46

erta

-0.45

POSITIVE LOGITS

 للمعارف

0.78

TOP

0.76

ngdoc

0.75

Portail

0.72

 tops

0.72

 Tops

0.71

 podium

0.71

ViewInit

0.70

XmlType

0.68

IntoConstraints

0.66

Activations Density 0.129%

references to rankings or lists that indicate top positions or achievements

No Comments

No Known Activations