INDEX

Explanations

the presence of the terms related to decoding and encoding mechanisms, particularly in the context of Protobuf

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 3-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_3/width_16k/average_l0_90

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.3.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

-0.46

 parten

-0.43

 DriverManager

-0.42

SBATCH

-0.41

{:.

-0.40

TPR

-0.40

 senhor

-0.38

izr

-0.38

 Impressive

-0.38

staff

-0.36

POSITIVE LOGITS

encode

2.39

decode

2.33

 encode

2.22

 decode

2.16

Encode

2.03

Decode

1.92

 Decode

1.85

 encoding

1.82

 encoded

1.80

 Encode

1.77

Activations Density 0.114%

the presence of the terms related to decoding and encoding mechanisms, particularly in the context of Protobuf

No Comments

No Known Activations