INDEX

Explanations

references to the past

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

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

hyde

-0.79

 Franch

-0.79

wagon

-0.72

NEY

-0.72

starting

-0.69

ithmetic

-0.66

anguage

-0.66

rolling

-0.65

roller

-0.65

mand

-0.65

POSITIVE LOGITS

ebin

1.69

decade

1.23

 tense

1.22

 week

0.95

orate

0.94

 century

0.93

 generations

0.91

ures

0.89

iche

0.89

 millennium

0.88

Activations Density 0.247%

references to the past

No Comments

No Known Activations