INDEX

Explanations

repeated conjunctions or phrases indicating addition or emphasis

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.20.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

 itſelf

-0.95

Efq

-0.93

 utafitiHapana

-0.85

 Majefty

-0.78

guien

-0.76

 Shakspeare

-0.76

ſelves

-0.73

 himſelf

-0.73

 themſelves

-0.71

 ingenieros

-0.69

POSITIVE LOGITS

1.46

rogens

1.22

 hence

1.17

 then

1.16

 thus

1.11

 therefore

1.09

romeda

1.05

rog

1.02

hence

0.99

ouille

0.98

Activations Density 1.400%

repeated conjunctions or phrases indicating addition or emphasis

No Comments

No Known Activations