INDEX

Explanations

references to protocol-related terms and structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 >=",

-0.88

ніципалі

-0.83

Eſ

-0.82

 Theſe

-0.81

Efq

-0.80

 NUKAT

-0.80

 MonoBehaviour

-0.79

 Reſ

-0.78

 ſtate

-0.77

 becauſe

-0.77

POSITIVE LOGITS

protobuf

0.77

 PROTO

0.75

 proto

0.69

 Proto

0.65

PROTO

0.52

proto

0.51

tipo

0.50

Proto

0.50

typal

0.49

typical

0.48

Activations Density 0.098%

references to protocol-related terms and structures

No Comments

No Known Activations