INDEX

Explanations

terms related to evaluation and interpretation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 5-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_5/width_16k/average_l0_68

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.5.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Alexei

-0.53

opcional

-0.52

romechanical

-0.52

optional

-0.51

Autoritní

-0.50

Ці

-0.48

 didat

-0.47

ilu

-0.47

rya

-0.47

 acum

-0.47

POSITIVE LOGITS

 regarded

0.99

 treated

0.87

 treating

0.86

awtextra

0.81

 Treated

0.81

Treating

0.79

setVerticalGroup

0.77

Treat

0.76

 Viewed

0.76

treat

0.75

Activations Density 0.344%

terms related to evaluation and interpretation

No Comments

No Known Activations