INDEX

Explanations

references to heads, whether literal or metaphorical

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 34-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_34/width_16k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.34.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

 OFDb

-0.48

DataMember

-0.48

 esternos

-0.48

FormBuilder

-0.44

EDEFAULT

-0.44

 unmute

-0.43

nextProps

-0.40

Cite

-0.40

 himo

-0.39

قایناق‌لار

-0.38

POSITIVE LOGITS

 head

2.98

 heads

2.73

 Head

2.48

Head

2.41

head

2.39

 HEAD

2.39

 Heads

2.27

HEAD

2.16

heads

2.16

 cabeza

2.13

Activations Density 0.569%

references to heads, whether literal or metaphorical

No Comments

No Known Activations