INDEX

Explanations

instances of the word "th" and variations of numerical references to court cases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 5-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_5/width_16k/average_l0_68

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.5.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

 newState

-0.66

❷

-0.62

 Brot

-0.58

oms

-0.55

anganan

-0.55

WART

-0.55

 Herzog

-0.55

 Bogen

-0.54

ATLAS

-0.54

 laranja

-0.54

POSITIVE LOGITS

1.46

 Five

1.30

⁵

1.25

 Fifth

1.24

 FIVE

1.21

five

1.17

Five

1.16

Fifth

1.15

cinq

1.12

FIVE

1.11

Activations Density 0.278%

instances of the word "th" and variations of numerical references to court cases

No Comments

No Known Activations