INDEX

Explanations

statements about actions and attributes of individuals, particularly focusing on their capabilities and experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 10-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_10/width_16k/average_l0_77

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.10.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

 Blech

-0.52

 luogo

-0.49

рс

-0.49

oad

-0.49

UVWXYZ

-0.48

Unione

-0.48

 yourselves

-0.48

 stalo

-0.48

 başlad

-0.47

jíma

-0.47

POSITIVE LOGITS

 himself

1.71

 Himself

1.36

himself

1.35

his

1.13

His

0.94

his

0.93

He

0.93

His

0.92

因为他

0.90

的他

0.90

Activations Density 0.647%

statements about actions and attributes of individuals, particularly focusing on their capabilities and experiences

No Comments

No Known Activations