INDEX

Explanations

references to entities, objects, and their interactions within a structured or code-like context

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

+#+#

-0.54

astéroïdes

-0.42

Décès

-0.40

Descripció

-0.36

Тем

-0.36

RTEE

-0.36

 comparison

-0.36

 CHAP

-0.35

erçe

-0.34

 Privacidade

-0.34

POSITIVE LOGITS

MLLoader

0.52

carteles

0.51

onAttach

0.48

StoryboardSegue

0.48

 endforeach

0.48

ValueGeneration

0.47

HasAnnotation

0.46

 utafitiHapana

0.46

 Baillargeon

0.45

andExpect

0.45

Activations Density 0.012%

references to entities, objects, and their interactions within a structured or code-like context

No Comments

No Known Activations