INDEX

Explanations

question phrases and the format of conversational exchanges

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

expandindo

-0.87

bootstrapcdn

-0.86

parsedMessage

-0.86

 zwiſchen

-0.85

 geſch

-0.82

tagHelperRunner

-0.79

<unused68>

-0.79

<unused28>

-0.79

[@BOS@]

-0.78

<unused3>

-0.78

POSITIVE LOGITS

What

0.46

 questions

0.42

How

0.41

 asked

0.40

Is

0.40

 What

0.39

 question

0.37

uestions

0.37

why

0.36

Why

0.36

Activations Density 0.304%

question phrases and the format of conversational exchanges

No Comments

No Known Activations