INDEX

Explanations

terms related to evaluation and judgment of competence or capability

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

之助

-0.38

nantes

-0.37

rel

-0.35

PREFIX

-0.35

Quarter

-0.34

ped

-0.34

readystatechange

-0.34

StartTag

-0.34

Prefix

-0.33

quoise

-0.33

POSITIVE LOGITS

rungsseite

0.65

Personendaten

0.61

 snippetHide

0.59

 AssemblyCompany

0.58

ViewFeatures

0.52

 الحره

0.52

脚注の使い方

0.52

\{\\

0.52

 برانيه

0.49

/\.

0.49

Activations Density 0.080%

terms related to evaluation and judgment of competence or capability

No Comments

No Known Activations