INDEX

Explanations

terms related to evaluation and assessment in various contexts, particularly scientific and legal

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

numbers and symbols

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

 المعيارى

-0.83

 ब्रेकडाउन

-0.79

 dieſen

-0.79

 Мексичка

-0.78

 queſta

-0.77

ſelben

-0.77

цездатний

-0.76

 dieſem

-0.75

 zuſammen

-0.75

 verſch

-0.75

POSITIVE LOGITS

setVerticalGroup

0.32

us

0.31

al

0.31

ec

0.30

ownic

0.30

0.29

0.28

The

0.28

Activations Density 0.081%

terms related to evaluation and assessment in various contexts, particularly scientific and legal

numbers and symbols

No Comments

No Known Activations