INDEX

Explanations

affirmations and acknowledgments in conversations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 17-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_17/width_16k/average_l0_73

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.17.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

 pillar

-0.42

 wealth

-0.40

 volume

-0.37

Cli

-0.37

DNC

-0.36

 Pillar

-0.36

 display

-0.35

gem

-0.35

 death

-0.35

HL

-0.34

POSITIVE LOGITS

Okay

0.97

 Okay

0.93

OK

0.90

 OKAY

0.88

OK

0.85

OKAY

0.83

 okay

0.83

Ok

0.82

okay

0.82

Okey

0.78

Activations Density 0.027%

affirmations and acknowledgments in conversations

No Comments

No Known Activations