INDEX

Explanations

terminology related to ranking and evaluation metrics

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

ranking, scores, priority

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

patrón

-0.47

AsUp

-0.47

 HasFactory

-0.47

ingale

-0.46

ulgação

-0.45

quila

-0.45

 galley

-0.45

Története

-0.44

gregation

-0.43

arache

-0.43

POSITIVE LOGITS

 ranking

0.58

 score

0.58

 rank

0.57

Rank

0.55

 scores

0.53

 Rank

0.50

 rankings

0.49

 ranked

0.47

twimg

0.46

rank

0.46

Activations Density 0.169%

terminology related to ranking and evaluation metrics

ranking, scores, priority

No Comments

No Known Activations