INDEX

Explanations

concepts related to holding or preserving items or data

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

PreExecute

-0.89

?></

-0.83

()?;

-0.83

 bomber

-0.79

Bow

-0.77

 Cowley

-0.77

++);

-0.77

PageInfo

-0.76

 bombers

-0.75

 Blumen

-0.75

POSITIVE LOGITS

 Holder

1.85

holder

1.79

 holder

1.78

 holders

1.67

holders

1.64

Holder

1.64

 Holders

1.61

HOLDER

1.57

ホルダー

1.12

HOLD

0.94

Activations Density 0.056%

concepts related to holding or preserving items or data

No Comments

No Known Activations