INDEX

Explanations

terms related to specific formats and quantities in various contexts

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

RegressionTest

-0.51

forName

-0.43

PushMatrix

-0.43

 officielle

-0.42

Hozzáférés

-0.42

 setempat

-0.41

 presidencia

-0.40

Cyfeiriadau

-0.40

harapkan

-0.40

 responsável

-0.39

POSITIVE LOGITS

RTGC

0.55

 EconPapers

0.53

 nakalista

0.52

RTDA

0.51

 PyTuple

0.47

CascadeType

0.43

 inverted

0.43

 singly

0.43

RTLU

0.43

ichio

0.42

Activations Density 0.060%

terms related to specific formats and quantities in various contexts

No Comments

No Known Activations