INDEX

Explanations

terms related to overriding configurations or settings

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

 fær

-0.51

Clik

-0.45

ulemon

-0.44

DATAS

-0.43

jsonwebtoken

-0.42

 للمعارف

-0.42

när

-0.40

IsContent

-0.39

Dato

-0.38

POSITIVE LOGITS

Override

1.12

 Override

1.09

override

1.08

verride

0.96

 override

0.94

 Overrides

0.88

overrides

0.82

 overrides

0.82

 overridden

0.80

Overrides

0.78

Activations Density 0.009%

terms related to overriding configurations or settings

No Comments

No Known Activations