INDEX

Explanations

metrics related to evaluation or assessment processes

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

another sentence starts

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

 nakalista

-0.79

Personendaten

-0.73

 Roskov

-0.68

SharedDtor

-0.64

ParallelGroup

-0.63

EDEFAULT

-0.61

 Савезне

-0.59

AddTagHelper

-0.57

SourceChecksum

-0.56

SequentialGroup

-0.56

POSITIVE LOGITS

vœ

0.45

 recours

0.36

numerusform

0.35

Enregistrer

0.35

 femininas

0.34

ientôt

0.34

 dignité

0.34

 nôtre

0.33

 lembran

0.33

 rendement

0.33

Activations Density 0.018%

metrics related to evaluation or assessment processes

another sentence starts

No Comments

No Known Activations