INDEX

Explanations

classifications and groupings based on various characteristics or patterns

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

phases, types, steps, modes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

########.

-0.64

 lenker

-0.52

 ComVisible

-0.51

 Wiktionnaire

-0.51

stdc

-0.49

reportWebVitals

-0.48

OGND

-0.47

+#+

-0.47

httphttps

-0.46

 beginnetje

-0.45

POSITIVE LOGITS

stateProvider

0.47

 types

0.39

StatusOK

0.37

envolvimento

0.36

babkan

0.36

potamus

0.36

();)

0.36

fallen

0.36

urlpatterns

0.35

Activations Density 0.122%

classifications and groupings based on various characteristics or patterns

phases, types, steps, modes

No Comments

No Known Activations