INDEX

Explanations

occurrences of specific logical or mathematical keywords and conditions

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

 pleaſure

-0.82

 faſt

-0.74

 ſche

-0.71

 Anſ

-0.71

 ſta

-0.70

MLLoader

-0.69

InstrumentedTest

-0.69

AddTagHelper

-0.66

 Geſ

-0.66

 enfans

-0.65

POSITIVE LOGITS

 among

0.48

 amongst

0.41

import

0.40

 AMONG

0.38

 Among

0.38

==

0.37

notin

0.36

among

0.36

 disponibilité

0.35

or

0.35

Activations Density 0.008%

occurrences of specific logical or mathematical keywords and conditions

No Comments

No Known Activations