INDEX

Explanations

characters and interactions in dialogues

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_23/width_131k/average_l0_101

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.23.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

 लेकर

-0.41

tierung

-0.39

 dekoratif

-0.39

 manguera

-0.38

 bordada

-0.37

 dramatique

-0.36

 typically

-0.35

 cremallera

-0.35

EndContext

-0.35

ができ

-0.35

POSITIVE LOGITS

writeFieldEnd

0.57

featureID

0.51

⟬

0.47

########.

0.47

AntiForgeryToken

0.45

hyrchwyd

0.45

 CallOverrides

0.44

 interlocutor

0.44

SPD

0.43

httphttps

0.43

Activations Density 0.032%

characters and interactions in dialogues

No Comments

No Known Activations