INDEX

Explanations

terms related to preferences or preferred types in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 17-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_17/width_65k/average_l0_125

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

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

 DialogInterface

-0.65

 المعيارى

-0.59

 Offers

-0.59

 OFFERS

-0.59

 invitations

-0.57

PRNewswire

-0.57

</thead>

-0.56

нской

-0.56

ExecuteReader

-0.54

lapse

-0.54

POSITIVE LOGITS

isTrue

0.56

љено

0.56

StoryboardSegue

0.55

 prochains

0.54

ième

0.54

ựng

0.52

SuppressMessage

0.50

 propOrder

0.50

::_('

0.50

期刊论文

0.50

Activations Density 0.105%

terms related to preferences or preferred types in various contexts

No Comments

No Known Activations