INDEX

Explanations

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

adh

-0.69

Published

-0.65

erer

-0.62

Daily

-0.62

 Beans

-0.61

oway

-0.59

Place

-0.59

Dispatch

-0.59

arden

-0.58

ribution

-0.58

POSITIVE LOGITS

lihood

1.25

 sized

1.04

 twins

1.03

 minded

0.96

worldly

0.91

 vein

0.90

minded

0.85

ities

0.83

MpServer

0.77

arios

0.76

Activations Density 0.441%

similar words

No Comments

No Known Activations