INDEX

Explanations

repetitive phrases and expressions indicating agreement or adding information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 7-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_7/width_16k/average_l0_69

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.7.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

 للمعارف

-0.92

antaranya

-0.69

 ✭✭

-0.67

tvguidetime

-0.65

 Shakspeare

-0.63

Rüyada

-0.62

>`;

-0.62

berdayakan

-0.61

obenz

-0.59

 Lohan

-0.58

POSITIVE LOGITS

 andererseits

0.63

 also

0.58

 equally

0.57

同时

0.54

同時に

0.53

 conversely

0.53

但也

0.53

 ALSO

0.51

 simultaneously

0.50

 aussi

0.49

Activations Density 0.143%

repetitive phrases and expressions indicating agreement or adding information

No Comments

No Known Activations