INDEX

Explanations

descriptors related to weight categories and upright postures

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

ibal

-0.96

-0.94

ophon

-0.86

erate

-0.79

leton

-0.78

icipated

-0.77

asant

-0.77

gage

-0.77

enei

-0.76

lee

-0.76

POSITIVE LOGITS

 EDITION

0.70

 mode

0.56

FIG

0.55

 poultry

0.55

 brackets

0.54

 flats

0.54

 clause

0.53

Wem

0.53

weed

0.53

LET

0.53

Activations Density 0.079%

descriptors related to weight categories and upright postures

No Comments

No Known Activations