INDEX

Explanations

season numbers and premiere information related to TV shows

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

heid

-0.75

hend

-0.71

aith

-0.71

plom

-0.69

rish

-0.67

yre

-0.67

zsche

-0.66

company

-0.66

amina

-0.65

orney

-0.65

POSITIVE LOGITS

 premiere

1.21

 Premiere

1.19

 Episode

1.19

 finale

1.12

 onwards

1.07

 onward

0.96

 episode

0.94

 premie

0.94

 episodes

0.92

 premier

0.84

Activations Density 0.026%

season numbers and premiere information related to TV shows

No Comments

No Known Activations