INDEX

Explanations

tokens representing programming concepts or language keywords

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

Mathematical outputs in computational text, particularly prime factorization results and numerical comparison answers that appear after specific question prompts.

eleuther_acts_top20 · claude-3-7-sonnet-20250219 Triggered by @lu-christina

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

RectangleBorder

-0.82

ſicht

-0.79

InputBorder

-0.79

__":

-0.79

addGap

-0.77

__':

-0.77

DockStyle

-0.77

__":

-0.77

ScopeManager

-0.74

OGND

-0.73

POSITIVE LOGITS

 écrits

0.33

 étrangère

0.33

0.31

 religión

0.30

 scolaires

0.30

 méri

0.30

 étrangères

0.30

 souhait

0.29

 gyhoeddwyd

0.29

 chré

0.29

Activations Density 0.105%

tokens representing programming concepts or language keywords

Mathematical outputs in computational text, particularly prime factorization results and numerical comparison answers that appear after specific question prompts.

No Comments

No Known Activations