INDEX

Explanations

instances of high numerical values or significant emphasis within the text

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

citations, numbers, and legal text

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

 feroit

-0.97

 pouvoit

-0.93

 plufieurs

-0.91

 auroit

-0.89

 queſta

-0.89

 ainfi

-0.84

 dezelve

-0.83

ésultats

-0.82

 étoient

-0.81

 ſch

-0.81

POSITIVE LOGITS

0.45

0.43

↵

0.42

↵↵

0.40

、

0.40

↵↵↵↵↵

0.38

↵↵↵↵

0.38

↵↵↵

0.38

0.37

Activations Density 0.108%

instances of high numerical values or significant emphasis within the text

citations, numbers, and legal text

No Comments

No Known Activations