INDEX

Explanations

elements related to instructions or guidance

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

Non-English text and code snippets

np_acts-logits-general · gemini-2.0-flash

ijerph, Lett. ijerph, Lett

np_acts-logits-general · gemini-2.5-flash-lite

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

 myſelf

-1.49

 itſelf

-1.42

 houſe

-1.38

 pleaſure

-1.34

 Monfieur

-1.33

Efq

-1.32

 themſelves

-1.30

 ſtate

-1.29

 Houſe

-1.28

 Reſ

-1.27

POSITIVE LOGITS

0.71

0.59

di

0.57

 einem

0.56

dem

0.54

old

0.54

0.53

Activations Density 0.025%

elements related to instructions or guidance

Non-English text and code snippets

ijerph, Lett. ijerph, Lett

No Comments

No Known Activations