INDEX

Explanations

references to competitive placements or rankings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

caster

-0.85

pher

-0.78

umen

-0.76

bek

-0.74

raltar

-0.73

metry

-0.68

mits

-0.66

 therap

-0.64

Unity

-0.64

worms

-0.61

POSITIVE LOGITS

 Century

1.02

 century

0.99

century

0.88

ablishment

0.83

 installment

0.77

 millennium

0.75

 edition

0.72

 Amendment

0.71

retch

0.68

 incarnation

0.67

Activations Density 0.008%

references to competitive placements or rankings

No Comments

No Known Activations