INDEX

Explanations

words related to actions or requests for help and assistance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 12-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_12/width_16k/average_l0_82

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.12.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Everybody

-0.71

 Everybody

-0.68

 girls

-0.63

 shrinks

-0.63

 talking

-0.62

 فريبيس

-0.61

big

-0.61

pij

-0.61

 people

-0.61

 thinks

-0.61

POSITIVE LOGITS

 erfolgte

0.83

此处

0.80

 herein

0.79

/−

0.79

 tevens

0.78

tadır

0.78

Дан

0.78

 attire

0.78

 paramInt

0.77

 kombat

0.77

Activations Density 1.711%

words related to actions or requests for help and assistance

No Comments

No Known Activations