INDEX

Explanations

dates and time-related terms

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

Reviewer

-0.91

 gorilla

-0.65

 louder

-0.61

 voic

-0.60

adium

-0.56

sers

-0.55

erers

-0.55

itcher

-0.54

icken

-0.54

oeuv

-0.54

POSITIVE LOGITS

 September

1.26

 January

1.24

 October

1.23

 June

1.22

 August

1.22

 July

1.21

 December

1.19

 February

1.18

 November

1.18

January

1.18

Activations Density 2.544%

dates and time-related terms

No Comments

No Known Activations