INDEX

Explanations

references to halls and venues, particularly those associated with fame or awards

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.4.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

 myſelf

-0.86

 Anſ

-0.86

 Perſ

-0.83

 ſind

-0.80

 Conſ

-0.80

 Reſ

-0.80

 Diſ

-0.79

 uſed

-0.75

 themſelves

-0.74

 uſe

-0.74

POSITIVE LOGITS

 hall

2.83

 Hall

2.79

 HALL

2.58

Hall

2.55

hall

2.40

 halls

2.34

HALL

2.25

 Halls

2.10

 hallway

1.38

 hallways

1.12

Activations Density 0.048%

references to halls and venues, particularly those associated with fame or awards

No Comments

No Known Activations