INDEX

Explanations

locations and their relevance to various contexts

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

specific parts or aspects

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

glBindBuffer

-0.50

 ExecuteAsync

-0.43

发表于

-0.42

httphttps

-0.41

]++;

-0.38

+/**

-0.36

WriteTagHelper

-0.35

]));

-0.35

thereum

-0.35

)>=

-0.35

POSITIVE LOGITS

ConstraintMaker

0.57

 Normdatei

0.50

CppMethod

0.50

kuuta

0.48

angkutan

0.48

 geleden

0.48

standigheden

0.48

tamment

0.47

 تعالى

0.47

 endroits

0.47

Activations Density 0.054%

locations and their relevance to various contexts

specific parts or aspects

No Comments

No Known Activations